code
stringlengths
17
6.64M
def generate_instruction_following_data(output_dir='./', seed_tasks_path='./seed_tasks.jsonl', num_instructions_to_generate=20000, model_name='text-davinci-003', num_prompt_instructions=3, request_batch_size=5, temperature=1.0, top_p=1.0, num_cpus=16): seed_tasks = [json.loads(l) for l in open(seed_tasks_path, 'r')] seed_instruction_data = [{'instruction': t['instruction'], 'input': t['instances'][0]['input'], 'output': t['instances'][0]['output']} for t in seed_tasks] print(f'Loaded {len(seed_instruction_data)} human-written seed instructions') os.makedirs(output_dir, exist_ok=True) request_idx = 0 machine_instruction_data = [] if os.path.exists(os.path.join(output_dir, 'regen.json')): machine_instruction_data = utils.jload(os.path.join(output_dir, 'regen.json')) print(f'Loaded {len(machine_instruction_data)} machine-generated instructions') scorer = rouge_scorer.RougeScorer(['rougeL'], use_stemmer=False) progress_bar = tqdm.tqdm(total=num_instructions_to_generate) if machine_instruction_data: progress_bar.update(len(machine_instruction_data)) all_instructions = ([d['instruction'] for d in seed_instruction_data] + [d['instruction'] for d in machine_instruction_data]) all_instruction_tokens = [scorer._tokenizer.tokenize(inst) for inst in all_instructions] while (len(machine_instruction_data) < num_instructions_to_generate): request_idx += 1 batch_inputs = [] for _ in range(request_batch_size): prompt_instructions = random.sample(seed_instruction_data, num_prompt_instructions) prompt = encode_prompt(prompt_instructions) batch_inputs.append(prompt) decoding_args = utils.OpenAIDecodingArguments(temperature=temperature, n=1, max_tokens=3072, top_p=top_p, stop=['\n20', '20.', '20.']) request_start = time.time() print('Calling openai...') results = utils.openai_completion(prompts=batch_inputs, model_name=model_name, batch_size=request_batch_size, decoding_args=decoding_args, logit_bias={'50256': (- 100)}) request_duration = (time.time() - request_start) print(f'request took - {request_duration}') process_start = time.time() instruction_data = [] for result in results: new_instructions = post_process_gpt3_response(num_prompt_instructions, result) instruction_data += new_instructions total = len(instruction_data) keep = 0 for instruction_data_entry in instruction_data: new_instruction_tokens = scorer._tokenizer.tokenize(instruction_data_entry['instruction']) with Pool(num_cpus) as p: rouge_scores = p.map(partial(rouge_scorer._score_lcs, new_instruction_tokens), all_instruction_tokens) rouge_scores = [score.fmeasure for score in rouge_scores] most_similar_instructions = {all_instructions[i]: rouge_scores[i] for i in np.argsort(rouge_scores)[(- 10):][::(- 1)]} if (max(rouge_scores) > 0.7): continue else: keep += 1 instruction_data_entry['most_similar_instructions'] = most_similar_instructions instruction_data_entry['avg_similarity_score'] = float(np.mean(rouge_scores)) machine_instruction_data.append(instruction_data_entry) all_instructions.append(instruction_data_entry['instruction']) all_instruction_tokens.append(new_instruction_tokens) progress_bar.update(1) process_duration = (time.time() - process_start) print(f'Request {request_idx} took {request_duration:.2f}s, processing took {process_duration:.2f}s') print(f'Generated {total} instructions, kept {keep} instructions') utils.jdump(machine_instruction_data, os.path.join(output_dir, 'regen.json'))
def main(task, **kwargs): globals()[task](**kwargs)
def train(base_model: str='', data_path: str='yahma/alpaca-cleaned', output_dir: str='./lora-alpaca', batch_size: int=128, micro_batch_size: int=8, num_epochs: int=1, learning_rate: float=0.0003, cutoff_len: int=2048, val_set_size: int=2000, lora_r: int=8, lora_alpha: int=16, lora_dropout: float=0.05, lora_target_modules: List[str]=['q_proj', 'v_proj'], train_on_inputs: bool=True, group_by_length: bool=False, wandb_project: str='', wandb_run_name: str='', wandb_watch: str='', wandb_log_model: str='', resume_from_checkpoint: str=None, prompt_template_name: str='alpaca'): if (int(os.environ.get('LOCAL_RANK', 0)) == 0): print(f'''Training Alpaca-LoRA model with params: base_model: {base_model} data_path: {data_path} output_dir: {output_dir} batch_size: {batch_size} micro_batch_size: {micro_batch_size} num_epochs: {num_epochs} learning_rate: {learning_rate} cutoff_len: {cutoff_len} val_set_size: {val_set_size} lora_r: {lora_r} lora_alpha: {lora_alpha} lora_dropout: {lora_dropout} lora_target_modules: {lora_target_modules} train_on_inputs: {train_on_inputs} group_by_length: {group_by_length} wandb_project: {wandb_project} wandb_run_name: {wandb_run_name} wandb_watch: {wandb_watch} wandb_log_model: {wandb_log_model} resume_from_checkpoint: {(resume_from_checkpoint or False)} prompt template: {prompt_template_name} ''') assert base_model, "Please specify a --base_model, e.g. --base_model='huggyllama/llama-7b'" gradient_accumulation_steps = 16 prompter = Prompter(prompt_template_name) device_map = 'auto' world_size = int(os.environ.get('WORLD_SIZE', 1)) ddp = (world_size != 1) if ddp: device_map = {'': int((os.environ.get('LOCAL_RANK') or 0))} gradient_accumulation_steps = (gradient_accumulation_steps // world_size) use_wandb = ((len(wandb_project) > 0) or (('WANDB_PROJECT' in os.environ) and (len(os.environ['WANDB_PROJECT']) > 0))) if (len(wandb_project) > 0): os.environ['WANDB_PROJECT'] = wandb_project if (len(wandb_watch) > 0): os.environ['WANDB_WATCH'] = wandb_watch if (len(wandb_log_model) > 0): os.environ['WANDB_LOG_MODEL'] = wandb_log_model model = AutoModelForCausalLM.from_pretrained(base_model, load_in_8bit=False, torch_dtype=torch.float16, device_map=device_map) tokenizer = AutoTokenizer.from_pretrained(base_model) tokenizer.pad_token_id = 0 tokenizer.padding_side = 'left' def tokenize(prompt, add_eos_token=True): result = tokenizer(prompt, truncation=True, max_length=cutoff_len, padding=False, return_tensors=None) if ((result['input_ids'][(- 1)] != tokenizer.eos_token_id) and (len(result['input_ids']) < cutoff_len) and add_eos_token): result['input_ids'].append(tokenizer.eos_token_id) result['attention_mask'].append(1) result['labels'] = result['input_ids'].copy() return result def generate_and_tokenize_prompt(data_point): full_prompt = prompter.generate_prompt(data_point['instruction'], data_point['input'], data_point['output']) tokenized_full_prompt = tokenize(full_prompt) if (not train_on_inputs): user_prompt = prompter.generate_prompt(data_point['instruction'], data_point['input']) tokenized_user_prompt = tokenize(user_prompt, add_eos_token=False) user_prompt_len = len(tokenized_user_prompt['input_ids']) tokenized_full_prompt['labels'] = (([(- 100)] * user_prompt_len) + tokenized_full_prompt['labels'][user_prompt_len:]) return tokenized_full_prompt if (data_path.endswith('.json') or data_path.endswith('.jsonl')): data = load_dataset('json', data_files=data_path) else: data = load_dataset(data_path) if resume_from_checkpoint: checkpoint_name = os.path.join(resume_from_checkpoint, 'pytorch_model.bin') if (not os.path.exists(checkpoint_name)): checkpoint_name = os.path.join(resume_from_checkpoint, 'adapter_model.bin') resume_from_checkpoint = False if os.path.exists(checkpoint_name): print(f'Restarting from {checkpoint_name}') adapters_weights = torch.load(checkpoint_name) else: print(f'Checkpoint {checkpoint_name} not found') if (val_set_size > 0): train_val = data['train'].train_test_split(test_size=val_set_size, shuffle=True, seed=42) train_data = train_val['train'].shuffle().map(generate_and_tokenize_prompt) val_data = train_val['test'].shuffle().map(generate_and_tokenize_prompt) else: train_data = data['train'].shuffle().map(generate_and_tokenize_prompt) val_data = None if ((not ddp) and (torch.cuda.device_count() > 1)): model.is_parallelizable = True model.model_parallel = True trainer = transformers.Trainer(model=model, train_dataset=train_data, eval_dataset=val_data, args=transformers.TrainingArguments(per_device_train_batch_size=batch_size, gradient_accumulation_steps=gradient_accumulation_steps, warmup_steps=100, num_train_epochs=num_epochs, learning_rate=learning_rate, fp16=True, logging_steps=10, optim='adamw_torch', evaluation_strategy=('steps' if (val_set_size > 0) else 'no'), save_strategy='steps', eval_steps=(200 if (val_set_size > 0) else None), save_steps=200, output_dir=output_dir, save_total_limit=3, load_best_model_at_end=(True if (val_set_size > 0) else False), ddp_find_unused_parameters=(False if ddp else None), report_to=('wandb' if use_wandb else None), run_name=(wandb_run_name if use_wandb else None)), data_collator=transformers.DataCollatorForSeq2Seq(tokenizer, pad_to_multiple_of=8, return_tensors='pt', padding=True)) model.config.use_cache = False if ((torch.__version__ >= '2') and (sys.platform != 'win32')): model = torch.compile(model) trainer.train(resume_from_checkpoint=resume_from_checkpoint) model.save_pretrained(output_dir) print("\n If there's a warning about missing keys above, please disregard :)")
@dataclass class ModelArguments(): model_name_or_path: Optional[str] = field(default='facebook/opt-125m')
@dataclass class DataArguments(): data_path: str = field(default=None, metadata={'help': 'Path to the training data.'})
@dataclass class TrainingArguments(transformers.TrainingArguments): cache_dir: Optional[str] = field(default=None) optim: str = field(default='adamw_torch') model_max_length: int = field(default=512, metadata={'help': 'Maximum sequence length. Sequences will be right padded (and possibly truncated).'})
def smart_tokenizer_and_embedding_resize(special_tokens_dict: Dict, tokenizer: transformers.PreTrainedTokenizer, model: transformers.PreTrainedModel): 'Resize tokenizer and embedding.\n\n Note: This is the unoptimized version that may make your embedding size not be divisible by 64.\n ' num_new_tokens = tokenizer.add_special_tokens(special_tokens_dict) model.resize_token_embeddings(len(tokenizer)) if (num_new_tokens > 0): input_embeddings = model.get_input_embeddings().weight.data output_embeddings = model.get_output_embeddings().weight.data input_embeddings_avg = input_embeddings[:(- num_new_tokens)].mean(dim=0, keepdim=True) output_embeddings_avg = output_embeddings[:(- num_new_tokens)].mean(dim=0, keepdim=True) input_embeddings[(- num_new_tokens):] = input_embeddings_avg output_embeddings[(- num_new_tokens):] = output_embeddings_avg
def _tokenize_fn(strings: Sequence[str], tokenizer: transformers.PreTrainedTokenizer) -> Dict: 'Tokenize a list of strings.' tokenized_list = [tokenizer(text, return_tensors='pt', padding='longest', max_length=tokenizer.model_max_length, truncation=True) for text in strings] input_ids = labels = [tokenized.input_ids[0] for tokenized in tokenized_list] input_ids_lens = labels_lens = [tokenized.input_ids.ne(tokenizer.pad_token_id).sum().item() for tokenized in tokenized_list] return dict(input_ids=input_ids, labels=labels, input_ids_lens=input_ids_lens, labels_lens=labels_lens)
def preprocess(sources: Sequence[str], targets: Sequence[str], tokenizer: transformers.PreTrainedTokenizer) -> Dict: 'Preprocess the data by tokenizing.' examples = [(s + t) for (s, t) in zip(sources, targets)] (examples_tokenized, sources_tokenized) = [_tokenize_fn(strings, tokenizer) for strings in (examples, sources)] input_ids = examples_tokenized['input_ids'] labels = copy.deepcopy(input_ids) for (label, source_len) in zip(labels, sources_tokenized['input_ids_lens']): label[:source_len] = IGNORE_INDEX return dict(input_ids=input_ids, labels=labels)
class SupervisedDataset(Dataset): 'Dataset for supervised fine-tuning.' def __init__(self, data_path: str, tokenizer: transformers.PreTrainedTokenizer): super(SupervisedDataset, self).__init__() logging.warning('Loading data...') list_data_dict = utils.jload(data_path) logging.warning('Formatting inputs...') (prompt_input, prompt_no_input) = (PROMPT_DICT['prompt_input'], PROMPT_DICT['prompt_no_input']) sources = [(prompt_input.format_map(example) if (example.get('input', '') != '') else prompt_no_input.format_map(example)) for example in list_data_dict] targets = [f"{example['output']}{tokenizer.eos_token}" for example in list_data_dict] logging.warning('Tokenizing inputs... This may take some time...') data_dict = preprocess(sources, targets, tokenizer) self.input_ids = data_dict['input_ids'] self.labels = data_dict['labels'] def __len__(self): return len(self.input_ids) def __getitem__(self, i) -> Dict[(str, torch.Tensor)]: return dict(input_ids=self.input_ids[i], labels=self.labels[i])
@dataclass class DataCollatorForSupervisedDataset(object): 'Collate examples for supervised fine-tuning.' tokenizer: transformers.PreTrainedTokenizer def __call__(self, instances: Sequence[Dict]) -> Dict[(str, torch.Tensor)]: (input_ids, labels) = tuple(([instance[key] for instance in instances] for key in ('input_ids', 'labels'))) input_ids = torch.nn.utils.rnn.pad_sequence(input_ids, batch_first=True, padding_value=self.tokenizer.pad_token_id) labels = torch.nn.utils.rnn.pad_sequence(labels, batch_first=True, padding_value=IGNORE_INDEX) return dict(input_ids=input_ids, labels=labels, attention_mask=input_ids.ne(self.tokenizer.pad_token_id))
def make_supervised_data_module(tokenizer: transformers.PreTrainedTokenizer, data_args) -> Dict: 'Make dataset and collator for supervised fine-tuning.' train_dataset = SupervisedDataset(tokenizer=tokenizer, data_path=data_args.data_path) data_collator = DataCollatorForSupervisedDataset(tokenizer=tokenizer) return dict(train_dataset=train_dataset, eval_dataset=None, data_collator=data_collator)
def train(): parser = transformers.HfArgumentParser((ModelArguments, DataArguments, TrainingArguments)) (model_args, data_args, training_args) = parser.parse_args_into_dataclasses() model = transformers.AutoModelForCausalLM.from_pretrained(model_args.model_name_or_path, cache_dir=training_args.cache_dir) tokenizer = transformers.AutoTokenizer.from_pretrained(model_args.model_name_or_path, cache_dir=training_args.cache_dir, model_max_length=training_args.model_max_length, padding_side='right', use_fast=False) if (tokenizer.pad_token is None): smart_tokenizer_and_embedding_resize(special_tokens_dict=dict(pad_token=DEFAULT_PAD_TOKEN), tokenizer=tokenizer, model=model) if ('llama' in model_args.model_name_or_path): tokenizer.add_special_tokens({'eos_token': DEFAULT_EOS_TOKEN, 'bos_token': DEFAULT_BOS_TOKEN, 'unk_token': DEFAULT_UNK_TOKEN}) data_module = make_supervised_data_module(tokenizer=tokenizer, data_args=data_args) trainer = Trainer(model=model, tokenizer=tokenizer, args=training_args, **data_module) trainer.train() trainer.save_model(training_args.output_dir)
@dataclasses.dataclass class OpenAIDecodingArguments(object): max_tokens: int = 1800 temperature: float = 0.2 top_p: float = 1.0 n: int = 1 stream: bool = False stop: Optional[Sequence[str]] = None presence_penalty: float = 0.0 frequency_penalty: float = 0.0 suffix: Optional[str] = None logprobs: Optional[int] = None echo: bool = False
def openai_completion(prompts: Union[(str, Sequence[str], Sequence[dict[(str, str)]], dict[(str, str)])], decoding_args: OpenAIDecodingArguments, model_name='text-davinci-003', sleep_time=2, batch_size=1, max_instances=sys.maxsize, max_batches=sys.maxsize, return_text=False, **decoding_kwargs) -> Union[(Union[StrOrOpenAIObject], Sequence[StrOrOpenAIObject], Sequence[Sequence[StrOrOpenAIObject]])]: 'Decode with OpenAI API.\n\n Args:\n prompts: A string or a list of strings to complete. If it is a chat model the strings should be formatted\n as explained here: https://github.com/openai/openai-python/blob/main/chatml.md. If it is a chat model\n it can also be a dictionary (or list thereof) as explained here:\n https://github.com/openai/openai-cookbook/blob/main/examples/How_to_format_inputs_to_ChatGPT_models.ipynb\n decoding_args: Decoding arguments.\n model_name: Model name. Can be either in the format of "org/model" or just "model".\n sleep_time: Time to sleep once the rate-limit is hit.\n batch_size: Number of prompts to send in a single request. Only for non chat model.\n max_instances: Maximum number of prompts to decode.\n max_batches: Maximum number of batches to decode. This argument will be deprecated in the future.\n return_text: If True, return text instead of full completion object (which contains things like logprob).\n decoding_kwargs: Additional decoding arguments. Pass in `best_of` and `logit_bias` if you need them.\n\n Returns:\n A completion or a list of completions.\n Depending on return_text, return_openai_object, and decoding_args.n, the completion type can be one of\n - a string (if return_text is True)\n - an openai_object.OpenAIObject object (if return_text is False)\n - a list of objects of the above types (if decoding_args.n > 1)\n ' is_single_prompt = isinstance(prompts, (str, dict)) if is_single_prompt: prompts = [prompts] if (max_batches < sys.maxsize): logging.warning('`max_batches` will be deprecated in the future, please use `max_instances` instead.Setting `max_instances` to `max_batches * batch_size` for now.') max_instances = (max_batches * batch_size) prompts = prompts[:max_instances] num_prompts = len(prompts) prompt_batches = [prompts[(batch_id * batch_size):((batch_id + 1) * batch_size)] for batch_id in range(int(math.ceil((num_prompts / batch_size))))] completions = [] for (batch_id, prompt_batch) in tqdm.tqdm(enumerate(prompt_batches), desc='prompt_batches', total=len(prompt_batches)): batch_decoding_args = copy.deepcopy(decoding_args) while True: try: shared_kwargs = dict(model=model_name, **batch_decoding_args.__dict__, **decoding_kwargs) completion_batch = openai.Completion.create(prompt=prompt_batch, **shared_kwargs) choices = completion_batch.choices for choice in choices: choice['total_tokens'] = completion_batch.usage.total_tokens completions.extend(choices) break except openai.error.OpenAIError as e: logging.warning(f'OpenAIError: {e}.') if ('Please reduce your prompt' in str(e)): batch_decoding_args.max_tokens = int((batch_decoding_args.max_tokens * 0.8)) logging.warning(f'Reducing target length to {batch_decoding_args.max_tokens}, Retrying...') else: logging.warning('Hit request rate limit; retrying...') time.sleep(sleep_time) if return_text: completions = [completion.text for completion in completions] if (decoding_args.n > 1): completions = [completions[i:(i + decoding_args.n)] for i in range(0, len(completions), decoding_args.n)] if is_single_prompt: (completions,) = completions return completions
def _make_w_io_base(f, mode: str): if (not isinstance(f, io.IOBase)): f_dirname = os.path.dirname(f) if (f_dirname != ''): os.makedirs(f_dirname, exist_ok=True) f = open(f, mode=mode) return f
def _make_r_io_base(f, mode: str): if (not isinstance(f, io.IOBase)): f = open(f, mode=mode) return f
def jdump(obj, f, mode='w', indent=4, default=str): 'Dump a str or dictionary to a file in json format.\n\n Args:\n obj: An object to be written.\n f: A string path to the location on disk.\n mode: Mode for opening the file.\n indent: Indent for storing json dictionaries.\n default: A function to handle non-serializable entries; defaults to `str`.\n ' f = _make_w_io_base(f, mode) if isinstance(obj, (dict, list)): json.dump(obj, f, indent=indent, default=default) elif isinstance(obj, str): f.write(obj) else: raise ValueError(f'Unexpected type: {type(obj)}') f.close()
def jload(f, mode='r'): 'Load a .json file into a dictionary.' f = _make_r_io_base(f, mode) jdict = json.load(f) f.close() return jdict
@register_model def mvit_tiny(pretrained=False, **kwargs): cfg = get_cfg() cfg_file = '../SlowFast_dev/configs/ImageNet/MVIT_T_10_CONV.yaml' cfg.merge_from_file(cfg_file) model = MViT(cfg) return model
class INatDataset(ImageFolder): def __init__(self, root, train=True, year=2018, transform=None, target_transform=None, category='name', loader=default_loader): self.transform = transform self.loader = loader self.target_transform = target_transform self.year = year path_json = os.path.join(root, f"{('train' if train else 'val')}{year}.json") with open(path_json) as json_file: data = json.load(json_file) with open(os.path.join(root, 'categories.json')) as json_file: data_catg = json.load(json_file) path_json_for_targeter = os.path.join(root, f'train{year}.json') with open(path_json_for_targeter) as json_file: data_for_targeter = json.load(json_file) targeter = {} indexer = 0 for elem in data_for_targeter['annotations']: king = [] king.append(data_catg[int(elem['category_id'])][category]) if (king[0] not in targeter.keys()): targeter[king[0]] = indexer indexer += 1 self.nb_classes = len(targeter) self.samples = [] for elem in data['images']: cut = elem['file_name'].split('/') target_current = int(cut[2]) path_current = os.path.join(root, cut[0], cut[2], cut[3]) categors = data_catg[target_current] target_current_true = targeter[categors[category]] self.samples.append((path_current, target_current_true))
def build_dataset(is_train, args): transform = build_transform(is_train, args) if (args.data_set == 'CIFAR'): dataset = datasets.CIFAR100(args.data_path, train=is_train, transform=transform) nb_classes = 100 elif (args.data_set == 'IMNET'): root = os.path.join(args.data_path, ('train' if is_train else 'val')) dataset = datasets.ImageFolder(root, transform=transform) nb_classes = 1000 elif (args.data_set == 'INAT'): dataset = INatDataset(args.data_path, train=is_train, year=2018, category=args.inat_category, transform=transform) nb_classes = dataset.nb_classes elif (args.data_set == 'INAT19'): dataset = INatDataset(args.data_path, train=is_train, year=2019, category=args.inat_category, transform=transform) nb_classes = dataset.nb_classes elif (args.data_set == 'IMNETH5'): dataset = Imagenet(args.data_path, train=is_train, transform=transform) nb_classes = 1000 return (dataset, nb_classes)
def build_transform(is_train, args): resize_im = (args.input_size > 32) if is_train: transform = create_transform(input_size=args.input_size, is_training=True, color_jitter=args.color_jitter, auto_augment=args.aa, interpolation=args.train_interpolation, re_prob=args.reprob, re_mode=args.remode, re_count=args.recount) if (not resize_im): transform.transforms[0] = transforms.RandomCrop(args.input_size, padding=4) return transform t = [] if resize_im: size = int(((256 / 224) * args.input_size)) t.append(transforms.Resize(size, interpolation=3)) t.append(transforms.CenterCrop(args.input_size)) t.append(transforms.ToTensor()) t.append(transforms.Normalize(IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD)) return transforms.Compose(t)
class Mlp(nn.Module): def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.0): super().__init__() out_features = (out_features or in_features) hidden_features = (hidden_features or in_features) self.fc1 = nn.Linear(in_features, hidden_features) self.act = act_layer() self.fc2 = nn.Linear(hidden_features, out_features) self.drop = nn.Dropout(drop) def forward(self, x): x = self.fc1(x) x = self.act(x) x = self.drop(x) x = self.fc2(x) x = self.drop(x) return x
class CMlp(nn.Module): def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.0): super().__init__() out_features = (out_features or in_features) hidden_features = (hidden_features or in_features) self.fc1 = nn.Conv2d(in_features, hidden_features, 1) self.act = act_layer() self.fc2 = nn.Conv2d(hidden_features, out_features, 1) self.drop = nn.Dropout(drop) def forward(self, x): x = self.fc1(x) x = self.act(x) x = self.drop(x) x = self.fc2(x) x = self.drop(x) return x
class GlobalSparseAttn(nn.Module): def __init__(self, dim, num_heads=8, qkv_bias=False, qk_scale=None, attn_drop=0.0, proj_drop=0.0, sr_ratio=1): super().__init__() self.num_heads = num_heads head_dim = (dim // num_heads) self.scale = (qk_scale or (head_dim ** (- 0.5))) self.qkv = nn.Linear(dim, (dim * 3), bias=qkv_bias) self.attn_drop = nn.Dropout(attn_drop) self.proj = nn.Linear(dim, dim) self.proj_drop = nn.Dropout(proj_drop) self.sr = sr_ratio if (self.sr > 1): self.sampler = nn.AvgPool2d(1, sr_ratio) kernel_size = sr_ratio self.LocalProp = nn.ConvTranspose2d(dim, dim, kernel_size, stride=sr_ratio, groups=dim) self.norm = nn.LayerNorm(dim) else: self.sampler = nn.Identity() self.upsample = nn.Identity() self.norm = nn.Identity() def forward(self, x, H: int, W: int): (B, N, C) = x.shape if (self.sr > 1.0): x = x.transpose(1, 2).reshape(B, C, H, W) x = self.sampler(x) x = x.flatten(2).transpose(1, 2) qkv = self.qkv(x).reshape(B, (- 1), 3, self.num_heads, (C // self.num_heads)).permute(2, 0, 3, 1, 4) (q, k, v) = (qkv[0], qkv[1], qkv[2]) attn = ((q @ k.transpose((- 2), (- 1))) * self.scale) attn = attn.softmax(dim=(- 1)) attn = self.attn_drop(attn) x = (attn @ v).transpose(1, 2).reshape(B, (- 1), C) if (self.sr > 1): x = x.permute(0, 2, 1).reshape(B, C, int((H / self.sr)), int((W / self.sr))) x = self.LocalProp(x) x = x.reshape(B, C, (- 1)).permute(0, 2, 1) x = self.norm(x) x = self.proj(x) x = self.proj_drop(x) return x
class LocalAgg(nn.Module): def __init__(self, dim, num_heads, mlp_ratio=4.0, qkv_bias=False, qk_scale=None, drop=0.0, attn_drop=0.0, drop_path=0.0, act_layer=nn.GELU, norm_layer=nn.LayerNorm): super().__init__() self.pos_embed = nn.Conv2d(dim, dim, 3, padding=1, groups=dim) self.norm1 = nn.BatchNorm2d(dim) self.conv1 = nn.Conv2d(dim, dim, 1) self.conv2 = nn.Conv2d(dim, dim, 1) self.attn = nn.Conv2d(dim, dim, 5, padding=2, groups=dim) self.drop_path = (DropPath(drop_path) if (drop_path > 0.0) else nn.Identity()) self.norm2 = nn.BatchNorm2d(dim) mlp_hidden_dim = int((dim * mlp_ratio)) self.mlp = CMlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop) def forward(self, x): x = (x + self.pos_embed(x)) x = (x + self.drop_path(self.conv2(self.attn(self.conv1(self.norm1(x)))))) x = (x + self.drop_path(self.mlp(self.norm2(x)))) return x
class SelfAttn(nn.Module): def __init__(self, dim, num_heads, mlp_ratio=4.0, qkv_bias=False, qk_scale=None, drop=0.0, attn_drop=0.0, drop_path=0.0, act_layer=nn.GELU, norm_layer=nn.LayerNorm, sr_ratio=1.0): super().__init__() self.pos_embed = nn.Conv2d(dim, dim, 3, padding=1, groups=dim) self.norm1 = norm_layer(dim) self.attn = GlobalSparseAttn(dim, num_heads=num_heads, qkv_bias=qkv_bias, qk_scale=qk_scale, attn_drop=attn_drop, proj_drop=drop, sr_ratio=sr_ratio) self.drop_path = (DropPath(drop_path) if (drop_path > 0.0) else nn.Identity()) self.norm2 = norm_layer(dim) mlp_hidden_dim = int((dim * mlp_ratio)) self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop) def forward(self, x): x = (x + self.pos_embed(x)) (B, N, H, W) = x.shape x = x.flatten(2).transpose(1, 2) x = (x + self.drop_path(self.attn(self.norm1(x), H, W))) x = (x + self.drop_path(self.mlp(self.norm2(x)))) x = x.transpose(1, 2).reshape(B, N, H, W) return x
class LGLBlock(nn.Module): def __init__(self, dim, num_heads, mlp_ratio=4.0, qkv_bias=False, qk_scale=None, drop=0.0, attn_drop=0.0, drop_path=0.0, act_layer=nn.GELU, norm_layer=nn.LayerNorm, sr_ratio=1.0): super().__init__() if (sr_ratio > 1): self.LocalAgg = LocalAgg(dim, num_heads, mlp_ratio, qkv_bias, qk_scale, drop, attn_drop, drop_path, act_layer, norm_layer) else: self.LocalAgg = nn.Identity() self.SelfAttn = SelfAttn(dim, num_heads, mlp_ratio, qkv_bias, qk_scale, drop, attn_drop, drop_path, act_layer, norm_layer, sr_ratio) def forward(self, x): x = self.LocalAgg(x) x = self.SelfAttn(x) return x
class PatchEmbed(nn.Module): ' Image to Patch Embedding\n ' def __init__(self, img_size=224, patch_size=16, in_chans=3, embed_dim=768): super().__init__() img_size = to_2tuple(img_size) patch_size = to_2tuple(patch_size) num_patches = ((img_size[1] // patch_size[1]) * (img_size[0] // patch_size[0])) self.img_size = img_size self.patch_size = patch_size self.num_patches = num_patches self.norm = nn.LayerNorm(embed_dim) self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size) def forward(self, x): (B, C, H, W) = x.shape assert ((H == self.img_size[0]) and (W == self.img_size[1])), f"Input image size ({H}*{W}) doesn't match model ({self.img_size[0]}*{self.img_size[1]})." x = self.proj(x) (B, C, H, W) = x.shape x = x.flatten(2).transpose(1, 2) x = self.norm(x) x = x.reshape(B, H, W, (- 1)).permute(0, 3, 1, 2).contiguous() return x
class EdgeVit(nn.Module): ' Vision Transformer\n A PyTorch impl of : `An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale` -\n https://arxiv.org/abs/2010.11929\n ' def __init__(self, depth=[1, 2, 5, 3], img_size=224, in_chans=3, num_classes=1000, embed_dim=[48, 96, 240, 384], head_dim=64, mlp_ratio=4.0, qkv_bias=True, qk_scale=None, representation_size=None, drop_rate=0.0, attn_drop_rate=0.0, drop_path_rate=0.0, norm_layer=None, sr_ratios=[4, 2, 2, 1], **kwargs): '\n Args:\n depth (list): depth of each stage\n img_size (int, tuple): input image size\n in_chans (int): number of input channels\n num_classes (int): number of classes for classification head\n embed_dim (list): embedding dimension of each stage\n head_dim (int): head dimension\n mlp_ratio (int): ratio of mlp hidden dim to embedding dim\n qkv_bias (bool): enable bias for qkv if True\n qk_scale (float): override default qk scale of head_dim ** -0.5 if set\n representation_size (Optional[int]): enable and set representation layer (pre-logits) to this value if set\n drop_rate (float): dropout rate\n attn_drop_rate (float): attention dropout rate\n drop_path_rate (float): stochastic depth rate\n norm_layer (nn.Module): normalization layer\n ' super().__init__() self.num_classes = num_classes self.num_features = self.embed_dim = embed_dim norm_layer = (norm_layer or partial(nn.LayerNorm, eps=1e-06)) self.patch_embed1 = PatchEmbed(img_size=img_size, patch_size=4, in_chans=in_chans, embed_dim=embed_dim[0]) self.patch_embed2 = PatchEmbed(img_size=(img_size // 4), patch_size=2, in_chans=embed_dim[0], embed_dim=embed_dim[1]) self.patch_embed3 = PatchEmbed(img_size=(img_size // 8), patch_size=2, in_chans=embed_dim[1], embed_dim=embed_dim[2]) self.patch_embed4 = PatchEmbed(img_size=(img_size // 16), patch_size=2, in_chans=embed_dim[2], embed_dim=embed_dim[3]) self.pos_drop = nn.Dropout(p=drop_rate) dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(depth))] num_heads = [(dim // head_dim) for dim in embed_dim] self.blocks1 = nn.ModuleList([LGLBlock(dim=embed_dim[0], num_heads=num_heads[0], mlp_ratio=mlp_ratio[0], qkv_bias=qkv_bias, qk_scale=qk_scale, drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i], norm_layer=norm_layer, sr_ratio=sr_ratios[0]) for i in range(depth[0])]) self.blocks2 = nn.ModuleList([LGLBlock(dim=embed_dim[1], num_heads=num_heads[1], mlp_ratio=mlp_ratio[1], qkv_bias=qkv_bias, qk_scale=qk_scale, drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[(i + depth[0])], norm_layer=norm_layer, sr_ratio=sr_ratios[1]) for i in range(depth[1])]) self.blocks3 = nn.ModuleList([LGLBlock(dim=embed_dim[2], num_heads=num_heads[2], mlp_ratio=mlp_ratio[2], qkv_bias=qkv_bias, qk_scale=qk_scale, drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[((i + depth[0]) + depth[1])], norm_layer=norm_layer, sr_ratio=sr_ratios[2]) for i in range(depth[2])]) self.blocks4 = nn.ModuleList([LGLBlock(dim=embed_dim[3], num_heads=num_heads[3], mlp_ratio=mlp_ratio[3], qkv_bias=qkv_bias, qk_scale=qk_scale, drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[(((i + depth[0]) + depth[1]) + depth[2])], norm_layer=norm_layer, sr_ratio=sr_ratios[3]) for i in range(depth[3])]) self.norm = nn.BatchNorm2d(embed_dim[(- 1)]) if representation_size: self.num_features = representation_size self.pre_logits = nn.Sequential(OrderedDict([('fc', nn.Linear(embed_dim, representation_size)), ('act', nn.Tanh())])) else: self.pre_logits = nn.Identity() self.head = (nn.Linear(embed_dim[(- 1)], num_classes) if (num_classes > 0) else nn.Identity()) self.apply(self._init_weights) def _init_weights(self, m): if isinstance(m, nn.Linear): trunc_normal_(m.weight, std=0.02) if (isinstance(m, nn.Linear) and (m.bias is not None)): nn.init.constant_(m.bias, 0) elif isinstance(m, nn.LayerNorm): nn.init.constant_(m.bias, 0) nn.init.constant_(m.weight, 1.0) @torch.jit.ignore def no_weight_decay(self): return {'pos_embed', 'cls_token'} def get_classifier(self): return self.head def reset_classifier(self, num_classes, global_pool=''): self.num_classes = num_classes self.head = (nn.Linear(self.embed_dim, num_classes) if (num_classes > 0) else nn.Identity()) def forward_features(self, x): x = self.patch_embed1(x) x = self.pos_drop(x) for blk in self.blocks1: x = blk(x) x = self.patch_embed2(x) for blk in self.blocks2: x = blk(x) x = self.patch_embed3(x) for blk in self.blocks3: x = blk(x) x = self.patch_embed4(x) for blk in self.blocks4: x = blk(x) x = self.norm(x) x = self.pre_logits(x) return x def forward(self, x): x = self.forward_features(x) x = x.flatten(2).mean((- 1)) x = self.head(x) return x
@register_model def edgevit_xxs(pretrained=True, **kwargs): model = EdgeVit(depth=[1, 1, 3, 2], embed_dim=[36, 72, 144, 288], head_dim=36, mlp_ratio=([4] * 4), qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-06), sr_ratios=[4, 2, 2, 1], **kwargs) model.default_cfg = _cfg() return model
@register_model def edgevit_xs(pretrained=True, **kwargs): model = EdgeVit(depth=[1, 1, 3, 1], embed_dim=[48, 96, 240, 384], head_dim=48, mlp_ratio=([4] * 4), qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-06), sr_ratios=[4, 2, 2, 1], **kwargs) model.default_cfg = _cfg() return model
@register_model def edgevit_s(pretrained=True, **kwargs): model = EdgeVit(depth=[1, 2, 5, 3], embed_dim=[48, 96, 240, 384], head_dim=48, mlp_ratio=([4] * 4), qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-06), sr_ratios=[4, 2, 2, 1], **kwargs) model.default_cfg = _cfg() return model
def train_one_epoch(model: torch.nn.Module, criterion: DistillationLoss, data_loader: Iterable, optimizer: torch.optim.Optimizer, device: torch.device, epoch: int, loss_scaler, max_norm: float=0, model_ema: Optional[ModelEma]=None, mixup_fn: Optional[Mixup]=None, set_training_mode=True): model.train(set_training_mode) metric_logger = utils.MetricLogger(delimiter=' ') metric_logger.add_meter('lr', utils.SmoothedValue(window_size=1, fmt='{value:.6f}')) header = 'Epoch: [{}]'.format(epoch) print_freq = 10 for (samples, targets) in metric_logger.log_every(data_loader, print_freq, header): samples = samples.to(device, non_blocking=True) targets = targets.to(device, non_blocking=True) if (mixup_fn is not None): (samples, targets) = mixup_fn(samples, targets) with torch.cuda.amp.autocast(): outputs = model(samples) loss = criterion(samples, outputs, targets) loss_value = loss.item() if (not math.isfinite(loss_value)): print('Loss is {}, stopping training'.format(loss_value)) sys.exit(1) optimizer.zero_grad() is_second_order = (hasattr(optimizer, 'is_second_order') and optimizer.is_second_order) loss_scaler(loss, optimizer, clip_grad=max_norm, parameters=model.parameters(), create_graph=is_second_order) torch.cuda.synchronize() if (model_ema is not None): model_ema.update(model) metric_logger.update(loss=loss_value) metric_logger.update(lr=optimizer.param_groups[0]['lr']) metric_logger.synchronize_between_processes() print('Averaged stats:', metric_logger) return {k: meter.global_avg for (k, meter) in metric_logger.meters.items()}
@torch.no_grad() def evaluate(data_loader, model, device): criterion = torch.nn.CrossEntropyLoss() metric_logger = utils.MetricLogger(delimiter=' ') header = 'Test:' model.eval() for (images, target) in metric_logger.log_every(data_loader, 10, header): images = images.to(device, non_blocking=True) target = target.to(device, non_blocking=True) with torch.cuda.amp.autocast(): output = model(images) loss = criterion(output, target) (acc1, acc5) = accuracy(output, target, topk=(1, 5)) batch_size = images.shape[0] metric_logger.update(loss=loss.item()) metric_logger.meters['acc1'].update(acc1.item(), n=batch_size) metric_logger.meters['acc5'].update(acc5.item(), n=batch_size) metric_logger.synchronize_between_processes() print('* Acc@1 {top1.global_avg:.3f} Acc@5 {top5.global_avg:.3f} loss {losses.global_avg:.3f}'.format(top1=metric_logger.acc1, top5=metric_logger.acc5, losses=metric_logger.loss)) return {k: meter.global_avg for (k, meter) in metric_logger.meters.items()}
class Imagenet(torch.utils.data.Dataset): 'ImageNet dataset.' def __init__(self, root_path, train=True, transform=None): if train: self.mode = 'train' else: self.mode = 'val' self.data_path = root_path self._construct_imdb_h5() self.transform = transform def safe_record_loader(self, raw_frame, attempts=10, retry_delay=1): for j in range(attempts): try: img = Image.open(io.BytesIO(raw_frame)).convert('RGB') return img except OSError as e: print(f'''Attempt {j}/{attempts}: failed to load {e}''', flush=True) if (j == (attempts - 1)): raise time.sleep(retry_delay) def _construct_imdb_h5(self): self.h5_fp = os.path.join(self.data_path, (self.mode + '.h5')) assert osp.isfile(self.h5_fp), 'File not found: {}'.format(self.h5_fp) self.h5_file = None h5_file = h5py.File(self.h5_fp, 'r') self._imdb = [] labels = sorted(list(h5_file.keys())) for (key, value) in h5_file.items(): target = labels.index(key) for img_name in value.keys(): self._imdb.append({'image_name': img_name, 'class_name': key, 'class': target}) self.num_videos = len(self._imdb) def __load_h5__(self, index): try: if (self.h5_file is None): self.h5_file = h5py.File(self.h5_fp, 'r') record = self._imdb[index] raw_frame = self.h5_file[record['class_name']][record['image_name']][()] img = self.safe_record_loader(raw_frame) return img except Exception: return None def __getitem__(self, index): im = self.__load_h5__(index) if (self.transform is not None): im = self.transform(im) label = self._imdb[index]['class'] return (im, label) def __len__(self): return len(self._imdb)
class DistillationLoss(torch.nn.Module): '\n This module wraps a standard criterion and adds an extra knowledge distillation loss by\n taking a teacher model prediction and using it as additional supervision.\n ' def __init__(self, base_criterion: torch.nn.Module, teacher_model: torch.nn.Module, distillation_type: str, alpha: float, tau: float): super().__init__() self.base_criterion = base_criterion self.teacher_model = teacher_model assert (distillation_type in ['none', 'soft', 'hard']) self.distillation_type = distillation_type self.alpha = alpha self.tau = tau def forward(self, inputs, outputs, labels): '\n Args:\n inputs: The original inputs that are feed to the teacher model\n outputs: the outputs of the model to be trained. It is expected to be\n either a Tensor, or a Tuple[Tensor, Tensor], with the original output\n in the first position and the distillation predictions as the second output\n labels: the labels for the base criterion\n ' outputs_kd = None if (not isinstance(outputs, torch.Tensor)): (outputs, outputs_kd) = outputs base_loss = self.base_criterion(outputs, labels) if (self.distillation_type == 'none'): return base_loss if (outputs_kd is None): raise ValueError('When knowledge distillation is enabled, the model is expected to return a Tuple[Tensor, Tensor] with the output of the class_token and the dist_token') with torch.no_grad(): teacher_outputs = self.teacher_model(inputs) if (self.distillation_type == 'soft'): T = self.tau distillation_loss = ((F.kl_div(F.log_softmax((outputs_kd / T), dim=1), F.log_softmax((teacher_outputs / T), dim=1), reduction='sum', log_target=True) * (T * T)) / outputs_kd.numel()) elif (self.distillation_type == 'hard'): distillation_loss = F.cross_entropy(outputs_kd, teacher_outputs.argmax(dim=1)) loss = ((base_loss * (1 - self.alpha)) + (distillation_loss * self.alpha)) return loss
class RASampler(torch.utils.data.Sampler): 'Sampler that restricts data loading to a subset of the dataset for distributed,\n with repeated augmentation.\n It ensures that different each augmented version of a sample will be visible to a\n different process (GPU)\n Heavily based on torch.utils.data.DistributedSampler\n ' def __init__(self, dataset, num_replicas=None, rank=None, shuffle=True, num_repeats: int=3): if (num_replicas is None): if (not dist.is_available()): raise RuntimeError('Requires distributed package to be available') num_replicas = dist.get_world_size() if (rank is None): if (not dist.is_available()): raise RuntimeError('Requires distributed package to be available') rank = dist.get_rank() if (num_repeats < 1): raise ValueError('num_repeats should be greater than 0') self.dataset = dataset self.num_replicas = num_replicas self.rank = rank self.num_repeats = num_repeats self.epoch = 0 self.num_samples = int(math.ceil(((len(self.dataset) * self.num_repeats) / self.num_replicas))) self.total_size = (self.num_samples * self.num_replicas) self.num_selected_samples = int(math.floor((((len(self.dataset) // 256) * 256) / self.num_replicas))) self.shuffle = shuffle def __iter__(self): if self.shuffle: g = torch.Generator() g.manual_seed(self.epoch) indices = torch.randperm(len(self.dataset), generator=g) else: indices = torch.arange(start=0, end=len(self.dataset)) indices = torch.repeat_interleave(indices, repeats=self.num_repeats, dim=0).tolist() padding_size: int = (self.total_size - len(indices)) if (padding_size > 0): indices += indices[:padding_size] assert (len(indices) == self.total_size) indices = indices[self.rank:self.total_size:self.num_replicas] assert (len(indices) == self.num_samples) return iter(indices[:self.num_selected_samples]) def __len__(self): return self.num_selected_samples def set_epoch(self, epoch): self.epoch = epoch
class SmoothedValue(object): 'Track a series of values and provide access to smoothed values over a\n window or the global series average.\n ' def __init__(self, window_size=20, fmt=None): if (fmt is None): fmt = '{median:.4f} ({global_avg:.4f})' self.deque = deque(maxlen=window_size) self.total = 0.0 self.count = 0 self.fmt = fmt def update(self, value, n=1): self.deque.append(value) self.count += n self.total += (value * n) def synchronize_between_processes(self): '\n Warning: does not synchronize the deque!\n ' if (not is_dist_avail_and_initialized()): return t = torch.tensor([self.count, self.total], dtype=torch.float64, device='cuda') dist.barrier() dist.all_reduce(t) t = t.tolist() self.count = int(t[0]) self.total = t[1] @property def median(self): d = torch.tensor(list(self.deque)) return d.median().item() @property def avg(self): d = torch.tensor(list(self.deque), dtype=torch.float32) return d.mean().item() @property def global_avg(self): return (self.total / self.count) @property def max(self): return max(self.deque) @property def value(self): return self.deque[(- 1)] def __str__(self): return self.fmt.format(median=self.median, avg=self.avg, global_avg=self.global_avg, max=self.max, value=self.value)
class MetricLogger(object): def __init__(self, delimiter='\t'): self.meters = defaultdict(SmoothedValue) self.delimiter = delimiter def update(self, **kwargs): for (k, v) in kwargs.items(): if isinstance(v, torch.Tensor): v = v.item() assert isinstance(v, (float, int)) self.meters[k].update(v) def __getattr__(self, attr): if (attr in self.meters): return self.meters[attr] if (attr in self.__dict__): return self.__dict__[attr] raise AttributeError("'{}' object has no attribute '{}'".format(type(self).__name__, attr)) def __str__(self): loss_str = [] for (name, meter) in self.meters.items(): loss_str.append('{}: {}'.format(name, str(meter))) return self.delimiter.join(loss_str) def synchronize_between_processes(self): for meter in self.meters.values(): meter.synchronize_between_processes() def add_meter(self, name, meter): self.meters[name] = meter def log_every(self, iterable, print_freq, header=None): i = 0 if (not header): header = '' start_time = time.time() end = time.time() iter_time = SmoothedValue(fmt='{avg:.4f}') data_time = SmoothedValue(fmt='{avg:.4f}') space_fmt = ((':' + str(len(str(len(iterable))))) + 'd') log_msg = [header, (('[{0' + space_fmt) + '}/{1}]'), 'eta: {eta}', '{meters}', 'time: {time}', 'data: {data}'] if torch.cuda.is_available(): log_msg.append('max mem: {memory:.0f}') log_msg = self.delimiter.join(log_msg) MB = (1024.0 * 1024.0) for obj in iterable: data_time.update((time.time() - end)) (yield obj) iter_time.update((time.time() - end)) if (((i % print_freq) == 0) or (i == (len(iterable) - 1))): eta_seconds = (iter_time.global_avg * (len(iterable) - i)) eta_string = str(datetime.timedelta(seconds=int(eta_seconds))) if torch.cuda.is_available(): print(log_msg.format(i, len(iterable), eta=eta_string, meters=str(self), time=str(iter_time), data=str(data_time), memory=(torch.cuda.max_memory_allocated() / MB))) else: print(log_msg.format(i, len(iterable), eta=eta_string, meters=str(self), time=str(iter_time), data=str(data_time))) i += 1 end = time.time() total_time = (time.time() - start_time) total_time_str = str(datetime.timedelta(seconds=int(total_time))) print('{} Total time: {} ({:.4f} s / it)'.format(header, total_time_str, (total_time / len(iterable))))
def _load_checkpoint_for_ema(model_ema, checkpoint): '\n Workaround for ModelEma._load_checkpoint to accept an already-loaded object\n ' mem_file = io.BytesIO() torch.save(checkpoint, mem_file) mem_file.seek(0) model_ema._load_checkpoint(mem_file)
def setup_for_distributed(is_master): '\n This function disables printing when not in master process\n ' import builtins as __builtin__ builtin_print = __builtin__.print def print(*args, **kwargs): force = kwargs.pop('force', False) if (is_master or force): builtin_print(*args, **kwargs) __builtin__.print = print
def is_dist_avail_and_initialized(): if (not dist.is_available()): return False if (not dist.is_initialized()): return False return True
def get_world_size(): if (not is_dist_avail_and_initialized()): return 1 return dist.get_world_size()
def get_rank(): if (not is_dist_avail_and_initialized()): return 0 return dist.get_rank()
def is_main_process(): return (get_rank() == 0)
def save_on_master(*args, **kwargs): if is_main_process(): torch.save(*args, **kwargs)
def init_distributed_mode(args): if (('RANK' in os.environ) and ('WORLD_SIZE' in os.environ)): args.rank = int(os.environ['RANK']) args.world_size = int(os.environ['WORLD_SIZE']) args.gpu = int(os.environ['LOCAL_RANK']) elif ('SLURM_PROCID' in os.environ): args.rank = int(os.environ['SLURM_PROCID']) args.gpu = (args.rank % torch.cuda.device_count()) else: print('Not using distributed mode') args.distributed = False return args.distributed = True torch.cuda.set_device(args.gpu) args.dist_backend = 'nccl' print('| distributed init (rank {}): {}'.format(args.rank, args.dist_url), flush=True) torch.distributed.init_process_group(backend=args.dist_backend, init_method=args.dist_url, world_size=args.world_size, rank=args.rank) torch.distributed.barrier() setup_for_distributed((args.rank == 0))
class BatchGraph(): def __init__(self): self.graph = DGLGraph() self.number_of_nodes = 0 self.graphid_to_nodeids = {} self.num_of_subgraphs = 0 def add_subgraph(self, _g): assert isinstance(_g, DGLGraph) num_new_nodes = _g.number_of_nodes() self.graphid_to_nodeids[self.num_of_subgraphs] = torch.LongTensor(list(range(self.number_of_nodes, (self.number_of_nodes + num_new_nodes)))) self.graph.add_nodes(num_new_nodes, data=_g.ndata) (sources, dests) = _g.all_edges() sources += self.number_of_nodes dests += self.number_of_nodes self.graph.add_edges(sources, dests, data=_g.edata) self.number_of_nodes += num_new_nodes self.num_of_subgraphs += 1 def cuda(self, device=None): for k in self.graphid_to_nodeids.keys(): self.graphid_to_nodeids[k] = self.graphid_to_nodeids[k].cuda(device=device) def de_batchify_graphs(self, features=None): if (features is None): features = self.graph.ndata['features'] assert isinstance(features, torch.Tensor) vectors = [features.index_select(dim=0, index=self.graphid_to_nodeids[gid]) for gid in self.graphid_to_nodeids.keys()] lengths = [f.size(0) for f in vectors] max_len = max(lengths) for (i, v) in enumerate(vectors): vectors[i] = torch.cat((v, torch.zeros(size=((max_len - v.size(0)), *v.shape[1:]), requires_grad=v.requires_grad, device=v.device)), dim=0) output_vectors = torch.stack(vectors) lengths = torch.LongTensor(lengths).to(device=output_vectors.device) return (output_vectors, lengths) def get_network_inputs(self, cuda=False): raise NotImplementedError('Must be implemented by subclasses.')
class GGNNBatchGraph(BatchGraph): def get_network_inputs(self, cuda=False, device=None): features = self.graph.ndata['features'] edge_types = self.graph.edata['etype'] if cuda: self.cuda(device=device) return (self.graph, features.cuda(device=device), edge_types.cuda(device=device)) else: return (self.graph, features, edge_types) pass
class DataEntry(): def __init__(self, datset, num_nodes, features, edges, target): self.dataset = datset self.num_nodes = num_nodes self.target = target self.graph = DGLGraph() self.features = torch.FloatTensor(features) self.graph.add_nodes(self.num_nodes, data={'features': self.features}) for (s, _type, t) in edges: etype_number = self.dataset.get_edge_type_number(_type) self.graph.add_edge(s, t, data={'etype': torch.LongTensor([etype_number])})
class DataSet(): def __init__(self, train_src, valid_src=None, test_src=None, batch_size=32, n_ident=None, g_ident=None, l_ident=None): self.train_examples = [] self.valid_examples = [] self.test_examples = [] self.train_batches = [] self.valid_batches = [] self.test_batches = [] self.batch_size = batch_size self.edge_types = {} self.max_etype = 0 self.feature_size = 0 (self.n_ident, self.g_ident, self.l_ident) = load_default_identifiers(n_ident, g_ident, l_ident) self.read_dataset(test_src, train_src, valid_src) self.initialize_dataset() def initialize_dataset(self): self.initialize_train_batch() self.initialize_valid_batch() self.initialize_test_batch() def read_dataset(self, test_src, train_src, valid_src): debug('Reading Train File!') with open(train_src) as fp: train_data = json.load(fp) for entry in tqdm(train_data): example = DataEntry(datset=self, num_nodes=len(entry[self.n_ident]), features=entry[self.n_ident], edges=entry[self.g_ident], target=entry[self.l_ident][0][0]) if (self.feature_size == 0): self.feature_size = example.features.size(1) debug(('Feature Size %d' % self.feature_size)) self.train_examples.append(example) if (valid_src is not None): debug('Reading Validation File!') with open(valid_src) as fp: valid_data = json.load(fp) for entry in tqdm(valid_data): example = DataEntry(datset=self, num_nodes=len(entry[self.n_ident]), features=entry[self.n_ident], edges=entry[self.g_ident], target=entry[self.l_ident][0][0]) self.valid_examples.append(example) if (test_src is not None): debug('Reading Test File!') with open(test_src) as fp: test_data = json.load(fp) for entry in tqdm(test_data): example = DataEntry(datset=self, num_nodes=len(entry[self.n_ident]), features=entry[self.n_ident], edges=entry[self.g_ident], target=entry[self.l_ident][0][0]) self.test_examples.append(example) def get_edge_type_number(self, _type): if (_type not in self.edge_types): self.edge_types[_type] = self.max_etype self.max_etype += 1 return self.edge_types[_type] @property def max_edge_type(self): return self.max_etype def initialize_train_batch(self, batch_size=(- 1)): if (batch_size == (- 1)): batch_size = self.batch_size self.train_batches = initialize_batch(self.train_examples, batch_size, shuffle=True) return len(self.train_batches) pass def initialize_valid_batch(self, batch_size=(- 1)): if (batch_size == (- 1)): batch_size = self.batch_size self.valid_batches = initialize_batch(self.valid_examples, batch_size) return len(self.valid_batches) pass def initialize_test_batch(self, batch_size=(- 1)): if (batch_size == (- 1)): batch_size = self.batch_size self.test_batches = initialize_batch(self.test_examples, batch_size) return len(self.test_batches) pass def get_dataset_by_ids_for_GGNN(self, entries, ids): taken_entries = [entries[i] for i in ids] labels = [e.target for e in taken_entries] batch_graph = GGNNBatchGraph() for entry in taken_entries: batch_graph.add_subgraph(copy.deepcopy(entry.graph)) return (batch_graph, torch.FloatTensor(labels)) def get_next_train_batch(self): if (len(self.train_batches) == 0): self.initialize_train_batch() ids = self.train_batches.pop() return self.get_dataset_by_ids_for_GGNN(self.train_examples, ids) def get_next_valid_batch(self): if (len(self.valid_batches) == 0): self.initialize_valid_batch() ids = self.valid_batches.pop() return self.get_dataset_by_ids_for_GGNN(self.valid_examples, ids) def get_next_test_batch(self): if (len(self.test_batches) == 0): self.initialize_test_batch() ids = self.test_batches.pop() return self.get_dataset_by_ids_for_GGNN(self.test_examples, ids)
class DevignModel(nn.Module): def __init__(self, input_dim, output_dim, max_edge_types, num_steps=8): super(DevignModel, self).__init__() self.inp_dim = input_dim self.out_dim = output_dim self.max_edge_types = max_edge_types self.num_timesteps = num_steps self.ggnn = GatedGraphConv(in_feats=input_dim, out_feats=output_dim, n_steps=num_steps, n_etypes=max_edge_types) self.conv_l1 = torch.nn.Conv1d(output_dim, output_dim, 3) self.maxpool1 = torch.nn.MaxPool1d(3, stride=2) self.conv_l2 = torch.nn.Conv1d(output_dim, output_dim, 1) self.maxpool2 = torch.nn.MaxPool1d(2, stride=2) self.concat_dim = (input_dim + output_dim) self.conv_l1_for_concat = torch.nn.Conv1d(self.concat_dim, self.concat_dim, 3) self.maxpool1_for_concat = torch.nn.MaxPool1d(3, stride=2) self.conv_l2_for_concat = torch.nn.Conv1d(self.concat_dim, self.concat_dim, 1) self.maxpool2_for_concat = torch.nn.MaxPool1d(2, stride=2) self.mlp_z = nn.Linear(in_features=self.concat_dim, out_features=1) self.mlp_y = nn.Linear(in_features=output_dim, out_features=1) self.sigmoid = nn.Sigmoid() def forward(self, batch, cuda=False): (graph, features, edge_types) = batch.get_network_inputs(cuda=cuda) outputs = self.ggnn(graph, features, edge_types) (x_i, _) = batch.de_batchify_graphs(features) (h_i, _) = batch.de_batchify_graphs(outputs) c_i = torch.cat((h_i, x_i), dim=(- 1)) (batch_size, num_node, _) = c_i.size() Y_1 = self.maxpool1(f.relu(self.conv_l1(h_i.transpose(1, 2)))) Y_2 = self.maxpool2(f.relu(self.conv_l2(Y_1))).transpose(1, 2) Z_1 = self.maxpool1_for_concat(f.relu(self.conv_l1_for_concat(c_i.transpose(1, 2)))) Z_2 = self.maxpool2_for_concat(f.relu(self.conv_l2_for_concat(Z_1))).transpose(1, 2) before_avg = torch.mul(self.mlp_y(Y_2), self.mlp_z(Z_2)) avg = before_avg.mean(dim=1) result = self.sigmoid(avg).squeeze(dim=(- 1)) return result
class GGNNSum(nn.Module): def __init__(self, input_dim, output_dim, max_edge_types, num_steps=8): super(GGNNSum, self).__init__() self.inp_dim = input_dim self.out_dim = output_dim self.max_edge_types = max_edge_types self.num_timesteps = num_steps self.ggnn = GatedGraphConv(in_feats=input_dim, out_feats=output_dim, n_steps=num_steps, n_etypes=max_edge_types) self.classifier = nn.Linear(in_features=output_dim, out_features=1) self.sigmoid = nn.Sigmoid() def forward(self, batch, cuda=False): (graph, features, edge_types) = batch.get_network_inputs(cuda=cuda) outputs = self.ggnn(graph, features, edge_types) (h_i, _) = batch.de_batchify_graphs(outputs) ggnn_sum = self.classifier(h_i.sum(dim=1)) result = self.sigmoid(ggnn_sum).squeeze(dim=(- 1)) return result
def evaluate_loss(model, loss_function, num_batches, data_iter, cuda=False): model.eval() with torch.no_grad(): _loss = [] (all_predictions, all_targets) = ([], []) for _ in range(num_batches): (graph, targets) = data_iter() targets = targets.cuda() predictions = model(graph, cuda=True) batch_loss = loss_function(predictions, targets) _loss.append(batch_loss.detach().cpu().item()) predictions = predictions.detach().cpu() if (predictions.ndim == 2): all_predictions.extend(np.argmax(predictions.numpy(), axis=(- 1)).tolist()) else: all_predictions.extend(predictions.ge(torch.ones(size=predictions.size()).fill_(0.5)).to(dtype=torch.int32).numpy().tolist()) all_targets.extend(targets.detach().cpu().numpy().tolist()) model.train() return (np.mean(_loss).item(), (accuracy_score(all_targets, all_predictions) * 100)) pass
def evaluate_metrics(model, loss_function, num_batches, data_iter): model.eval() with torch.no_grad(): _loss = [] (all_predictions, all_targets) = ([], []) for _ in range(num_batches): (graph, targets) = data_iter() targets = targets.cuda() predictions = model(graph, cuda=True) batch_loss = loss_function(predictions, targets) _loss.append(batch_loss.detach().cpu().item()) predictions = predictions.detach().cpu() if (predictions.ndim == 2): all_predictions.extend(np.argmax(predictions.numpy(), axis=(- 1)).tolist()) else: all_predictions.extend(predictions.ge(torch.ones(size=predictions.size()).fill_(0.5)).to(dtype=torch.int32).numpy().tolist()) all_targets.extend(targets.detach().cpu().numpy().tolist()) model.train() return ((accuracy_score(all_targets, all_predictions) * 100), (precision_score(all_targets, all_predictions) * 100), (recall_score(all_targets, all_predictions) * 100), (f1_score(all_targets, all_predictions) * 100)) pass
def train(model, dataset, max_steps, dev_every, loss_function, optimizer, save_path, log_every=50, max_patience=5): debug('Start Training') train_losses = [] best_model = None patience_counter = 0 best_f1 = 0 try: for step_count in range(max_steps): model.train() model.zero_grad() (graph, targets) = dataset.get_next_train_batch() targets = targets.cuda() predictions = model(graph, cuda=True) batch_loss = loss_function(predictions, targets) if ((log_every is not None) and ((step_count % log_every) == (log_every - 1))): debug(('Step %d\t\tTrain Loss %10.3f' % (step_count, batch_loss.detach().cpu().item()))) train_losses.append(batch_loss.detach().cpu().item()) batch_loss.backward() optimizer.step() if ((step_count % dev_every) == (dev_every - 1)): (valid_loss, valid_f1) = evaluate_loss(model, loss_function, dataset.initialize_train_batch(), dataset.get_next_train_batch) if (valid_f1 > best_f1): patience_counter = 0 best_f1 = valid_f1 best_model = copy.deepcopy(model.state_dict()) _save_file = open((save_path + '-model.bin'), 'wb') torch.save(model.state_dict(), _save_file) _save_file.close() else: patience_counter += 1 debug(('Step %d\t\tTrain Loss %10.3f\tValid Loss%10.3f\tf1: %5.2f\tPatience %d' % (step_count, np.mean(train_losses).item(), valid_loss, valid_f1, patience_counter))) debug(('=' * 100)) train_losses = [] if (patience_counter == max_patience): break except KeyboardInterrupt: debug('Training Interrupted by user!') if (best_model is not None): model.load_state_dict(best_model) _save_file = open((save_path + '-model.bin'), 'wb') torch.save(model.state_dict(), _save_file) _save_file.close() (acc, pr, rc, f1) = evaluate_metrics(model, loss_function, dataset.initialize_train_batch(), dataset.get_next_train_batch) debug(('%s\tTest Accuracy: %0.2f\tPrecision: %0.2f\tRecall: %0.2f\tF1: %0.2f' % (save_path, acc, pr, rc, f1))) debug(('=' * 100))
def load_default_identifiers(n, g, l): if (n is None): n = n_identifier if (g is None): g = g_identifier if (l is None): l = l_identifier return (n, g, l)
def initialize_batch(entries, batch_size, shuffle=False): total = len(entries) indices = np.arange(0, (total - 1), 1) if shuffle: np.random.shuffle(indices) batch_indices = [] start = 0 end = len(indices) curr = start while (curr < end): c_end = (curr + batch_size) if (c_end > end): c_end = end batch_indices.append(indices[curr:c_end]) curr = c_end return batch_indices[::(- 1)]
def tally_param(model): total = 0 for param in model.parameters(): total += param.data.nelement() return total
def debug(*msg, sep='\t'): caller = inspect.stack()[1] file_name = caller.filename ln = caller.lineno now = datetime.now() time = now.strftime('%m/%d/%Y - %H:%M:%S') print((((((('[' + str(time)) + '] File "') + file_name) + '", line ') + str(ln)) + ' '), end='\t') for m in msg: print(m, end=sep) print('')
class CSharpProcessor(): @classmethod def create_dead_for_loop(cls, body): control_variable = ('_i_' + str(np.random.choice(list(range(10))))) p = np.random.uniform(0, 1) if (p < 0.5): prefix = (((((('for ( int ' + control_variable) + ' = 0 ; ') + control_variable) + ' > 0 ; ') + control_variable) + ' ++ ) { ') loop = ((prefix + body) + ' } ') return loop else: return (('for ( ; false ; ) { ' + body) + '}') @classmethod def create_dead_while_loop(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('while ( false ) { ' + body) + ' }') elif (p < 0.66): return (((((('while ( ' + control_variable) + ' < ') + control_variable) + ' ) { ') + body) + ' } ') else: return (((((('while ( ' + control_variable) + ' > ') + control_variable) + ' ) { ') + body) + ' } ') @classmethod def create_dead_if(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('if ( false ) { ' + body) + ' }') elif (p < 0.66): return (((((('if ( ' + control_variable) + ' < ') + control_variable) + ' ) { ') + body) + ' } ') else: return (((((('if ( ' + control_variable) + ' > ') + control_variable) + ' ) { ') + body) + ' } ') @classmethod def for_to_while_random(cls, code_string, parser): root = parser.parse_code(code_string) loops = cls.extract_for_loops(root) success = False try: while ((not success) and (len(loops) > 0)): selected_loop = np.random.choice(loops) loops.remove(selected_loop) (modified_root, modified_code_string, success) = cls.for_to_while(code_string, root, selected_loop, parser) if success: root = modified_root code_string = modified_code_string except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_string, root)) return (root, code_string, success) @classmethod def while_to_for_random(cls, code_string, parser): root = parser.parse_code(code_string) loops = cls.extract_while_loops(root) success = False try: while ((not success) and (len(loops) > 0)): selected_loop = np.random.choice(loops) loops.remove(selected_loop) (modified_root, modified_code_string, success) = cls.while_to_for(code_string, root, selected_loop, parser) if success: root = modified_root code_string = modified_code_string if (not success): code_string = cls.beautify_java_code(get_tokens(code_string, root)) except: pass return (root, code_string, success) @classmethod def extract_for_loops(cls, root): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'for_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def beautify_java_code(cls, tokens): code = ' '.join(tokens) code = re.sub(' \\. ', '', code) code = re.sub(' \\+\\+', '++', code) return code @classmethod def get_tokens_replace_for(cls, code_str, for_node, root, init, cond, update, body): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode()) for child in children: if (child == for_node): tokens.extend(((((((init + [';', 'while', '(']) + cond) + [')', '{']) + body) + update) + ['}'])) else: tokens += cls.get_tokens_replace_for(code_str, for_node, child, init, cond, update, body) return tokens @classmethod def for_to_while(cls, code_string, root, fl, parser): children = fl.children init = children[2] init_tokens = get_tokens(code_string, init) comparison = children[4] if (str(comparison.type) != ';'): comp_tokens = get_tokens(code_string, comparison) update = children[6] if (str(update.type) == ')'): update_tokens = [] body = children[7] else: update_tokens = (get_tokens(code_string, update) + [';']) body = children[8] breaking_statements = cls.get_breaking_statements(body) body_tokens = cls.get_tokens_insert_before(code_string, body, ' '.join(update_tokens), breaking_statements) if ((len(body_tokens) >= 2) and ((body_tokens[0] == '{') and (body_tokens[(- 1)] == '}'))): body_tokens = body_tokens[1:(- 1)] tokens = cls.get_tokens_replace_for(code_str=code_string, for_node=fl, root=root, init=init_tokens, cond=comp_tokens, update=update_tokens, body=body_tokens) code = cls.beautify_java_code(tokens) return (parser.parse_code(code), code, True) return (root, code_string, False) @classmethod def get_tokens_insert_before(cls, code_str, root, insertion_code, insert_before_node): if (not isinstance(insert_before_node, list)): insert_before_node = [insert_before_node] if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): parent = root.parent if (len(parent.children) == 1): return tokens else: return [code_str[root.start_byte:root.end_byte].decode()] if (root in insert_before_node): tokens += insertion_code.split() children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode()) for child in children: ts = cls.get_tokens_insert_before(code_str, child, insertion_code, insert_before_node) tokens += ts return tokens @classmethod def extract_while_loops(cls, root): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'while_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def while_to_for(cls, code_string, root, wl, parser): children = wl.children condition = children[2] body = children[4] if (str(condition.type) == 'binary_expression'): expr_tokens = get_tokens(code_string, condition) body_tokens = get_tokens(code_string, body) if ((len(body_tokens) >= 2) and ((body_tokens[0] == '{') and (body_tokens[(- 1)] == '}'))): body_tokens = body_tokens[1:(- 1)] tokens = cls.get_tokens_replace_while(code_str=code_string, while_node=wl, root=root, cond=expr_tokens, body=body_tokens) code = cls.beautify_java_code(tokens) return (parser.parse_code(code), code, True) return (root, code_string, False) @classmethod def get_tokens_replace_while(cls, code_str, while_node, root, cond, body): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode()) for child in children: if (child == while_node): tokens.extend(((((['for', '(', ';'] + cond) + [';', ')', '{']) + body) + ['}'])) else: tokens += cls.get_tokens_replace_while(code_str, while_node, child, cond, body) return tokens @classmethod def extract_expression(self, root, code): expressions = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'binary_expression'): children_nodes = current_node.children keep = ['<', '>', '<=', '>=', '==', '!='] counter = 0 for w in children_nodes: if (str(w.type) in keep): counter = (counter + 1) if (counter == 1): expressions.append(current_node) for child in current_node.children: queue.append(child) return expressions @classmethod def get_tokens_for_opswap(cls, code, root, left_oprd, operator, right_oprd): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == operator.start_byte) and (root.end_byte == operator.end_byte)): opt = code[operator.start_byte:operator.end_byte].decode() if (opt == '<'): tokens.append('>') elif (opt == '>'): tokens.append('<') elif (opt == '>='): tokens.append('<=') elif (opt == '<='): tokens.append('>=') elif (opt == '=='): tokens.append('==') elif (opt == '!='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: if ((child.start_byte == left_oprd.start_byte) and (child.end_byte == left_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, right_oprd, left_oprd, operator, right_oprd) elif ((child.start_byte == right_oprd.start_byte) and (child.end_byte == right_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, left_oprd, left_oprd, operator, right_oprd) else: (ts, _) = cls.get_tokens_for_opswap(code, child, left_oprd, operator, right_oprd) tokens += ts return (tokens, None) @classmethod def operand_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) expressions = cls.extract_expression(root, code) success = False try: while ((not success) and (len(expressions) > 0)): selected_exp = np.random.choice(expressions) expressions.remove(selected_exp) bin_exp = selected_exp condition = code[bin_exp.start_byte:bin_exp.end_byte].decode() bin_exp = bin_exp.children left_oprd = bin_exp[0] operator = bin_exp[1] right_oprd = bin_exp[2] try: code_list = cls.get_tokens_for_opswap(code, root, left_oprd, operator, right_oprd)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success) @classmethod def extract_if_else(cls, root, code_str, operator_list): ext_opt_list = ['&&', '&', '||', '|'] expressions = [] queue = [root] not_consider = [] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'if_statement'): clause = code_str[current_node.start_byte:current_node.end_byte].decode() des = current_node.children[2] cond = code_str[des.start_byte:des.end_byte].decode() stack = [des] nodes = [] while (len(stack) > 0): root1 = stack.pop() if (len(root1.children) == 0): nodes.append(root1) for child in root1.children: stack.append(child) nodes.reverse() counter = 0 extra_counter = 0 for w in nodes: if (str(w.type) in operator_list): counter = (counter + 1) if (str(w.type) in ext_opt_list): extra_counter = (extra_counter + 1) if (not ((counter == 1) and (extra_counter == 0))): continue children_nodes = current_node.children flagx = 0 flagy = 0 for w in children_nodes: if (str(w.type) == 'else'): flagx = 1 if (str(w.type) == 'if_statement'): not_consider.append(w) flagy = 1 if ((flagx == 1) and (flagy == 0)): expressions.append([current_node, des]) for child in current_node.children: if (child not in not_consider): queue.append(child) return expressions @classmethod def get_tokens_for_blockswap(cls, code, root, first_block, opt_node, second_block, flagx, flagy): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == opt_node.start_byte) and (root.end_byte == opt_node.end_byte)): op = code[root.start_byte:root.end_byte].decode() if (op == '<'): tokens.append('>=') elif (op == '>'): tokens.append('<=') elif (op == '>='): tokens.append('<') elif (op == '<='): tokens.append('>') elif (op == '!='): tokens.append('==') elif (op == '=='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if ((child.start_byte == first_block.start_byte) and (child.end_byte == first_block.end_byte) and (flagx == 0) and (str(child.type) == str(first_block.type))): flagx = 1 (ts, _) = cls.get_tokens_for_blockswap(code, second_block, first_block, opt_node, second_block, flagx, flagy) elif ((child.start_byte == second_block.start_byte) and (child.end_byte == second_block.end_byte) and (flagy == 0) and (str(child.type) == str(second_block.type))): flagy = 1 (ts, _) = cls.get_tokens_for_blockswap(code, first_block, first_block, opt_node, second_block, flagx, flagy) else: (ts, _) = cls.get_tokens_for_blockswap(code, child, first_block, opt_node, second_block, flagx, flagy) tokens += ts return (tokens, None) @classmethod def block_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) operator_list = ['<', '>', '<=', '>=', '==', '!='] pair = cls.extract_if_else(root, code, operator_list) success = False lst = list(range(0, len(pair))) try: while ((not success) and (len(lst) > 0)): selected = np.random.choice(lst) lst.remove(selected) clause = pair[selected][0] des = pair[selected][1] st = [des] nodes = [] while (len(st) > 0): root1 = st.pop() if (len(root1.children) == 0): nodes.append(root1) if (code[root1.start_byte:root1.end_byte].decode() in operator_list): opt_node = root1 break for child in root1.children: st.append(child) nodes = clause.children flag = 0 for current_node in nodes: if (str(current_node.type) == 'block'): if (flag == 0): first_block = current_node flag = 1 else: second_block = current_node flagx = 0 flagy = 0 try: code_list = cls.get_tokens_for_blockswap(code, root, first_block, opt_node, second_block, flagx, flagy)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success) @classmethod def get_breaking_statements(cls, block): breakings = ['continue_statement', 'break_statement', 'return_statement'] statements = [] stack = [block] while (len(stack) > 0): top = stack.pop() if (str(top.type) in breakings): statements.append(top) else: for child in top.children: stack.append(child) return statements
class GoProcessor(): @classmethod def create_dead_for_loop(cls, body): control_variable = ('_i_' + str(np.random.choice(list(range(10))))) return (((f'for {control_variable} := 0 ; {control_variable} < 0; {control_variable}++' + ' { ') + body) + ' } ') @classmethod def create_dead_while_loop(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (((f'for false ' + ' { ') + body) + ' } ') elif (p < 0.66): return (((f'for {control_variable} > {control_variable} ' + ' { ') + body) + ' } ') else: return (((f'for {control_variable} < {control_variable} ' + ' { ') + body) + ' } ') @classmethod def create_dead_if(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (((f'if false ' + ' { ') + body) + ' } ') elif (p < 0.66): return (((f'if {control_variable} > {control_variable} ' + ' { ') + body) + ' } ') else: return (((f'if {control_variable} < {control_variable} ' + ' { ') + body) + ' } ') @classmethod def extract_expression(self, root, code): expressions = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'binary_expression'): children_nodes = current_node.children keep = ['<', '>', '<=', '>=', '==', '!='] counter = 0 for w in children_nodes: if (str(w.type) in keep): counter = (counter + 1) if (counter == 1): expressions.append(current_node) for child in current_node.children: queue.append(child) return expressions @classmethod def get_tokens_for_opswap(cls, code, root, left_oprd, operator, right_oprd): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == operator.start_byte) and (root.end_byte == operator.end_byte)): opt = code[operator.start_byte:operator.end_byte].decode() if (opt == '<'): tokens.append('>') elif (opt == '>'): tokens.append('<') elif (opt == '>='): tokens.append('<=') elif (opt == '<='): tokens.append('>=') elif (opt == '=='): tokens.append('==') elif (opt == '!='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: if ((child.start_byte == left_oprd.start_byte) and (child.end_byte == left_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, right_oprd, left_oprd, operator, right_oprd) elif ((child.start_byte == right_oprd.start_byte) and (child.end_byte == right_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, left_oprd, left_oprd, operator, right_oprd) else: (ts, _) = cls.get_tokens_for_opswap(code, child, left_oprd, operator, right_oprd) tokens += ts return (tokens, None) @classmethod def operand_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) expressions = cls.extract_expression(root, code) success = False try: while ((not success) and (len(expressions) > 0)): selected_exp = np.random.choice(expressions) expressions.remove(selected_exp) bin_exp = selected_exp condition = code[bin_exp.start_byte:bin_exp.end_byte].decode() bin_exp = bin_exp.children left_oprd = bin_exp[0] operator = bin_exp[1] right_oprd = bin_exp[2] try: code_list = cls.get_tokens_for_opswap(code, root, left_oprd, operator, right_oprd)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = get_tokens(code_str, root) return (code_string, success) @classmethod def extract_if_else(cls, root, code_str, operator_list): ext_opt_list = ['&&', '&', '||', '|'] expressions = [] queue = [root] not_consider = [] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'if_statement'): clause = code_str[current_node.start_byte:current_node.end_byte].decode() des = current_node.children[1] cond = code_str[des.start_byte:des.end_byte].decode() stack = [des] nodes = [] while (len(stack) > 0): root1 = stack.pop() if (len(root1.children) == 0): nodes.append(root1) for child in root1.children: stack.append(child) nodes.reverse() counter = 0 extra_counter = 0 for w in nodes: if (str(w.type) in operator_list): counter = (counter + 1) if (str(w.type) in ext_opt_list): extra_counter = (extra_counter + 1) if (not ((counter == 1) and (extra_counter == 0))): continue children_nodes = current_node.children flagx = 0 flagy = 0 for w in children_nodes: if (str(w.type) == 'else'): flagx = 1 if (str(w.type) == 'if_statement'): not_consider.append(w) flagy = 1 if ((flagx == 1) and (flagy == 0)): expressions.append([current_node, des]) for child in current_node.children: if (child not in not_consider): queue.append(child) return expressions @classmethod def get_tokens_for_blockswap(cls, code, root, first_block, opt_node, second_block, flagx, flagy): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == opt_node.start_byte) and (root.end_byte == opt_node.end_byte)): op = code[root.start_byte:root.end_byte].decode() if (op == '<'): tokens.append('>=') elif (op == '>'): tokens.append('<=') elif (op == '>='): tokens.append('<') elif (op == '<='): tokens.append('>') elif (op == '!='): tokens.append('==') elif (op == '=='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if ((child.start_byte == first_block.start_byte) and (child.end_byte == first_block.end_byte) and (flagx == 0) and (str(child.type) == str(first_block.type))): flagx = 1 (ts, _) = cls.get_tokens_for_blockswap(code, second_block, first_block, opt_node, second_block, flagx, flagy) elif ((child.start_byte == second_block.start_byte) and (child.end_byte == second_block.end_byte) and (flagy == 0) and (str(child.type) == str(second_block.type))): flagy = 1 (ts, _) = cls.get_tokens_for_blockswap(code, first_block, first_block, opt_node, second_block, flagx, flagy) else: (ts, _) = cls.get_tokens_for_blockswap(code, child, first_block, opt_node, second_block, flagx, flagy) tokens += ts return (tokens, None) @classmethod def block_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) operator_list = ['<', '>', '<=', '>=', '==', '!='] pair = cls.extract_if_else(root, code, operator_list) success = False lst = list(range(0, len(pair))) try: while ((not success) and (len(lst) > 0)): selected = np.random.choice(lst) lst.remove(selected) clause = pair[selected][0] des = pair[selected][1] st = [des] nodes = [] while (len(st) > 0): root1 = st.pop() if (len(root1.children) == 0): nodes.append(root1) if (code[root1.start_byte:root1.end_byte].decode() in operator_list): opt_node = root1 break for child in root1.children: st.append(child) nodes = clause.children flag = 0 for current_node in nodes: if (str(current_node.type) == 'block'): if (flag == 0): first_block = current_node flag = 1 else: second_block = current_node flagx = 0 flagy = 0 try: code_list = cls.get_tokens_for_blockswap(code, root, first_block, opt_node, second_block, flagx, flagy)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success) @classmethod def beautify_java_code(cls, tokens): code = ' '.join(tokens) code = re.sub(' \\. ', '', code) code = re.sub(' \\+\\+', '++', code) return code
class JavaAndCPPProcessor(): @classmethod def create_dead_for_loop(cls, body): control_variable = ('_i_' + str(np.random.choice(list(range(10))))) p = np.random.uniform(0, 1) if (p < 0.5): prefix = (((((('for ( int ' + control_variable) + ' = 0 ; ') + control_variable) + ' > 0 ; ') + control_variable) + ' ++ ) { ') loop = ((prefix + body) + ' } ') return loop else: return (('for ( ; false ; ) { ' + body) + '}') @classmethod def create_dead_while_loop(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('while ( false ) { ' + body) + ' }') elif (p < 0.66): return (((((('while ( ' + control_variable) + ' < ') + control_variable) + ' ) { ') + body) + ' } ') else: return (((((('while ( ' + control_variable) + ' > ') + control_variable) + ' ) { ') + body) + ' } ') @classmethod def create_dead_if(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('if ( false ) { ' + body) + ' }') elif (p < 0.66): return (((((('if ( ' + control_variable) + ' < ') + control_variable) + ' ) { ') + body) + ' } ') else: return (((((('if ( ' + control_variable) + ' > ') + control_variable) + ' ) { ') + body) + ' } ') @classmethod def for_to_while_random(cls, code_string, parser): root = parser.parse_code(code_string) loops = cls.extract_for_loops(root) success = False try: while ((not success) and (len(loops) > 0)): selected_loop = np.random.choice(loops) loops.remove(selected_loop) (modified_root, modified_code_string, success) = JavaAndCPPProcessor.for_to_while(code_string, root, selected_loop, parser) if success: root = modified_root code_string = modified_code_string except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_string, root)) return (root, code_string, success) @classmethod def while_to_for_random(cls, code_string, parser): root = parser.parse_code(code_string) loops = cls.extract_while_loops(root) success = False try: while ((not success) and (len(loops) > 0)): selected_loop = np.random.choice(loops) loops.remove(selected_loop) (modified_root, modified_code_string, success) = JavaAndCPPProcessor.while_to_for(code_string, root, selected_loop, parser) if success: root = modified_root code_string = modified_code_string if (not success): code_string = cls.beautify_java_code(get_tokens(code_string, root)) except: pass return (root, code_string, success) @classmethod def extract_for_loops(cls, root): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'for_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def beautify_java_code(cls, tokens): code = ' '.join(tokens) code = re.sub(' \\. ', '', code) code = re.sub(' \\+\\+', '++', code) return code @classmethod def get_tokens_replace_for(cls, code_str, for_node, root, init, cond, update, body): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode()) for child in children: if (child == for_node): tokens.extend(((((((init + ['while', '(']) + cond) + [')', '{']) + body) + update) + ['}'])) else: tokens += JavaAndCPPProcessor.get_tokens_replace_for(code_str, for_node, child, init, cond, update, body) return tokens @classmethod def extract_for_contents(cls, for_loop, code_string): children = for_loop.children init_part = children[2] if str(init_part.type).endswith('expression'): next_part_start = 4 init_tokens = (get_tokens(code_string, init_part) + [';']) elif (str(init_part.type).endswith('statement') or str(init_part.type).endswith('declaration')): next_part_start = 3 init_tokens = get_tokens(code_string, init_part) else: next_part_start = 3 init_tokens = [] comp_part = children[next_part_start] if str(comp_part.type).endswith('expression'): next_part_start += 2 comp_tokens = get_tokens(code_string, comp_part) else: comp_tokens = ['true'] next_part_start += 1 update_part = children[next_part_start] if str(update_part.type).endswith('expression'): next_part_start += 2 update_tokens = (get_tokens(code_string, update_part) + [';']) else: update_tokens = [] next_part_start += 1 block_part = children[next_part_start] breaking_statements = cls.get_breaking_statements(block_part) block_tokens = cls.get_tokens_insert_before(code_string, block_part, ' '.join(update_tokens), breaking_statements) return (init_tokens, comp_tokens, update_tokens, block_tokens) @classmethod def get_tokens_insert_before(cls, code_str, root, insertion_code, insert_before_node): if (not isinstance(insert_before_node, list)): insert_before_node = [insert_before_node] if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): parent = root.parent if (len(parent.children) == 1): return tokens else: return [code_str[root.start_byte:root.end_byte].decode()] if (root in insert_before_node): tokens += insertion_code.split() children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode()) for child in children: ts = cls.get_tokens_insert_before(code_str, child, insertion_code, insert_before_node) tokens += ts return tokens @classmethod def get_breaking_statements(cls, block): breakings = ['continue_statement', 'break_statement', 'return_statement'] statements = [] stack = [block] while (len(stack) > 0): top = stack.pop() if (str(top.type) in breakings): statements.append(top) else: for child in top.children: stack.append(child) return statements @classmethod def for_to_while(cls, code_string, root, fl, parser): original_tokenized_code = ' '.join(get_tokens(code_string, root)) (init_tokens, comp_tokens, update_tokens, body_tokens) = cls.extract_for_contents(fl, code_string) if ((len(body_tokens) >= 2) and ((body_tokens[0] == '{') and (body_tokens[(- 1)] == '}'))): body_tokens = body_tokens[1:(- 1)] tokens = cls.get_tokens_replace_for(code_str=code_string, for_node=fl, root=root, init=init_tokens, cond=comp_tokens, update=update_tokens, body=body_tokens) if (original_tokenized_code == ' '.join(tokens)): return (root, original_tokenized_code, False) code = cls.beautify_java_code(tokens) return (parser.parse_code(code), code, True) @classmethod def extract_while_loops(cls, root): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'while_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def while_to_for(cls, code_string, root, wl, parser): children = wl.children condition = children[1] body = children[2] if (str(condition.type) == 'parenthesized_expression'): expr_tokens = get_tokens(code_string, condition.children[1]) body_tokens = get_tokens(code_string, body) if ((len(body_tokens) >= 2) and ((body_tokens[0] == '{') and (body_tokens[(- 1)] == '}'))): body_tokens = body_tokens[1:(- 1)] tokens = cls.get_tokens_replace_while(code_str=code_string, while_node=wl, root=root, cond=expr_tokens, body=body_tokens) code = cls.beautify_java_code(tokens) return (parser.parse_code(code), code, True) return (root, code_string, False) @classmethod def get_tokens_replace_while(cls, code_str, while_node, root, cond, body): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode()) for child in children: if (child == while_node): tokens.extend(((((['for', '(', ';'] + cond) + [';', ')', '{']) + body) + ['}'])) else: tokens += JavaAndCPPProcessor.get_tokens_replace_while(code_str, while_node, child, cond, body) return tokens @classmethod def conditional_removal(cls, code_string, parser): root = parser.parse_code(code_string) (assi_cond_expr, varde_cond_expr, ret_cond_expr) = cls.extract_conditional_expression(root) success = False if (len(assi_cond_expr) > 0): try: modified_tokens = cls.assignment_conditional_removal(code_string, assi_cond_expr, root, parser) code_string = cls.beautify_java_code(modified_tokens) root = parser.parse_code(code_string) success = True (_, varde_cond_expr, ret_cond_expr) = cls.extract_conditional_expression(root) except: pass if (len(varde_cond_expr) > 0): try: modified_tokens = cls.var_decl_ternary_removal(code_string, varde_cond_expr, root, parser) code_string = cls.beautify_java_code(modified_tokens) root = parser.parse_code(code_string) success = True (_, _, ret_cond_expr) = cls.extract_conditional_expression(root) except: pass if (len(ret_cond_expr) > 0): try: modified_tokens = cls.return_ternary_removal(code_string, ret_cond_expr, root, parser) code_string = cls.beautify_java_code(modified_tokens) root = parser.parse_code(code_string) success = True except: pass return (root, code_string, success) @classmethod def assignment_conditional_removal(cls, code_string, assi_tern_expr, root, parser): if isinstance(code_string, str): code_string = code_string.encode() assert isinstance(root, Node) tokens = [] children = root.children if (len(children) == 0): tokens.append(cls.handle_terminal_node(root, code_string)) for child in children: if (child in assi_tern_expr): if (str(child.children[0].type) == 'conditional_expression'): cond_children = child.children[0].children if (str(cond_children[0].type) == 'assignment_expression'): assignee_token = get_tokens(code_string, cond_children[0].children[0])[0] condition_tokens = get_tokens(code_string, cond_children[0].children[2]) if (str(cond_children[0].children[2].type) == 'parenthesized_expression'): condition_tokens = condition_tokens[1:(- 1)] br1_tokens = get_tokens(code_string, cond_children[2]) br2_tokens = get_tokens(code_string, cond_children[4]) tokens.extend(((((((['if', '('] + condition_tokens) + [')', '{', assignee_token, '=']) + br1_tokens) + [';', '}', 'else', '{', assignee_token, '=']) + br2_tokens) + [';', '}'])) else: tokens += JavaAndCPPProcessor.assignment_conditional_removal(code_string, assi_tern_expr, child, parser) return tokens @classmethod def extract_conditional_expression(cls, root): assi_con_expr = [] varde_con_expr = [] ret_con_expr = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if ((str(current_node.type) == 'conditional_expression') and (str(current_node.children[0].type) == 'assignment_expression')): assi_con_expr.append(current_node.parent) if ((str(current_node.type) == 'conditional_expression') and (str(current_node.parent.type) == 'init_declarator')): varde_con_expr.append(current_node.parent.parent) if ((str(current_node.type) == 'conditional_expression') and (str(current_node.parent.type) == 'return_statement')): ret_con_expr.append(current_node.parent) for child in current_node.children: queue.append(child) return (assi_con_expr, varde_con_expr, ret_con_expr) @classmethod def ternary_removal(cls, code_string, parser): code_string = cls.remove_package_and_import(code_string) root = parser.parse_code(code_string) (assi_tern_expr, varde_tern_expr, ret_tern_expr) = cls.extract_ternary_expression(root) success = False if (len(assi_tern_expr) > 0): try: modified_tokens = cls.assignment_ternary_removal(code_string, assi_tern_expr, root, parser) code_string = cls.beautify_java_code(modified_tokens) root = parser.parse_code(code_string) success = True (_, varde_tern_expr, ret_tern_expr) = cls.extract_ternary_expression(root) except: pass if (len(varde_tern_expr) > 0): try: modified_tokens = cls.var_decl_ternary_removal(code_string, varde_tern_expr, root, parser) code_string = cls.beautify_java_code(modified_tokens) root = parser.parse_code(code_string) success = True (_, _, ret_tern_expr) = cls.extract_ternary_expression(root) except: pass if (len(ret_tern_expr) > 0): try: modified_tokens = cls.return_ternary_removal(code_string, ret_tern_expr, root, parser) code_string = cls.beautify_java_code(modified_tokens) root = parser.parse_code(code_string) success = True except: pass return (root, code_string, success) @classmethod def ternary_body_write(cls, body, code_string, assignee, tokens, ret=False): body_children = body.children condition_tokens = get_tokens(code_string, body_children[0]) if (str(body_children[0].type) == 'parenthesized_expression'): condition_tokens = condition_tokens[1:(- 1)] br1_tokens = get_tokens(code_string, body_children[2]) if (str(body_children[2].type) == 'parenthesized_expression'): br1_tokens = br1_tokens[1:(- 1)] br2_tokens = get_tokens(code_string, body_children[4]) if (str(body_children[4].type) == 'parenthesized_expression'): br2_tokens = br2_tokens[1:(- 1)] assignee_token = get_tokens(code_string, assignee)[0] if ret: tokens.extend(((((((['if', '('] + condition_tokens) + [')', '{', assignee_token]) + br1_tokens) + [';', '}', 'else', '{', assignee_token]) + br2_tokens) + [';', '}'])) else: tokens.extend(((((((['if', '('] + condition_tokens) + [')', '{', assignee_token, '=']) + br1_tokens) + [';', '}', 'else', '{', assignee_token, '=']) + br2_tokens) + [';', '}'])) return tokens @classmethod def assignment_ternary_removal(cls, code_string, assi_tern_expr, root, parser): if isinstance(code_string, str): code_string = code_string.encode() assert isinstance(root, Node) tokens = [] children = root.children if (len(children) == 0): tokens.append(cls.handle_terminal_node(root, code_string)) for child in children: if (child in assi_tern_expr): te_children = child.children assignee = te_children[0] body = te_children[2] tokens = cls.ternary_body_write(body, code_string, assignee, tokens) break else: tokens += JavaAndCPPProcessor.assignment_ternary_removal(code_string, assi_tern_expr, child, parser) return tokens @classmethod def var_decl_ternary_removal(cls, code_string, var_decl_tern_expr, root, parser): if isinstance(code_string, str): code_string = code_string.encode() assert isinstance(root, Node) tokens = [] children = root.children if (len(children) == 0): tokens.append(cls.handle_terminal_node(root, code_string)) for child in children: if (child in var_decl_tern_expr): for c in child.children: if (str(c.type) == ';'): continue elif ((str(c.type) == 'variable_declarator') or (str(c.type) == 'init_declarator')): te_children = c.children assignee = te_children[0] assignee_token = get_tokens(code_string, assignee)[0] tokens.extend([assignee_token, ';']) body = te_children[2] tokens = cls.ternary_body_write(body, code_string, assignee, tokens) else: tokens += get_tokens(code_string, c) else: tokens += JavaAndCPPProcessor.var_decl_ternary_removal(code_string, var_decl_tern_expr, child, parser) return tokens @classmethod def return_ternary_removal(cls, code_string, ret_tern_expr, root, parser): if isinstance(code_string, str): code_string = code_string.encode() assert isinstance(root, Node) tokens = [] children = root.children if (len(children) == 0): tokens.append(cls.handle_terminal_node(root, code_string)) for child in children: if (child in ret_tern_expr): te_children = child.children assignee = te_children[0] body = te_children[1] tokens = cls.ternary_body_write(body, code_string, assignee, tokens, ret=True) break else: tokens += JavaAndCPPProcessor.return_ternary_removal(code_string, ret_tern_expr, child, parser) return tokens @classmethod def extract_ternary_expression(cls, root): assi_ten_expr = [] varde_ten_expr = [] ret_ten_expr = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if ((str(current_node.type) == 'ternary_expression') and (str(current_node.parent.type) == 'assignment_expression')): assi_ten_expr.append(current_node.parent) if ((str(current_node.type) == 'ternary_expression') and (str(current_node.parent.type) == 'variable_declarator')): varde_ten_expr.append(current_node.parent.parent) if ((str(current_node.type) == 'ternary_expression') and (str(current_node.parent.type) == 'return_statement')): ret_ten_expr.append(current_node.parent) for child in current_node.children: queue.append(child) return (assi_ten_expr, varde_ten_expr, ret_ten_expr) @classmethod def incre_decre_removal(cls, code_string, parser): root = parser.parse_code(code_string) (pre_expr, post_expr) = cls.extract_incre_decre_expression(root, code_string) success = False if (len(pre_expr) > 0): try: modified_tokens = cls.pre_incre_decre_removal(code_string, pre_expr, root, parser) code_string = cls.beautify_java_code(modified_tokens) root = parser.parse_code(code_string) success = True (_, post_expr) = cls.extract_incre_decre_expression(root, code_string) except: pass if (len(post_expr) > 0): try: modified_tokens = cls.post_incre_decre_removal(code_string, post_expr, root, parser) code_string = cls.beautify_java_code(modified_tokens) root = parser.parse_code(code_string) success = True except: pass return (root, code_string, success) @classmethod def pre_incre_decre_removal(cls, code_string, pre_expr, root, parser): if isinstance(code_string, str): code_string = code_string.encode() assert isinstance(root, Node) tokens = [] children = root.children if (len(children) == 0): tokens.append(cls.handle_terminal_node(root, code_string)) for child in children: if (child in pre_expr): expr = child.children[0] assignee = expr.children[0] assignee_token = get_tokens(code_string, assignee)[0] op = '' if (str(expr.children[2].children[0].type) == '--'): op = '-=' elif (str(expr.children[2].children[0].type) == '++'): op = '+=' assigner = expr.children[2].children[(- 1)] assigner_token = get_tokens(code_string, assigner)[0] tokens.extend([assigner_token, op, '1', ';', assignee_token, '=', assigner_token, ';']) else: tokens += JavaAndCPPProcessor.pre_incre_decre_removal(code_string, pre_expr, child, parser) return tokens @classmethod def post_incre_decre_removal(cls, code_string, post_expr, root, parser): if isinstance(code_string, str): code_string = code_string.encode() assert isinstance(root, Node) tokens = [] children = root.children if (len(children) == 0): tokens.append(cls.handle_terminal_node(root, code_string)) for child in children: if (child in post_expr): expr = child.children[0] assignee = expr.children[0] assignee_token = get_tokens(code_string, assignee)[0] op = '' if (str(expr.children[2].children[(- 1)].type) == '--'): op = '-=' elif (str(expr.children[2].children[(- 1)].type) == '++'): op = '+=' assigner = expr.children[2].children[0] assigner_token = get_tokens(code_string, assigner)[0] tokens.extend([assignee_token, '=', assigner_token, ';', assigner_token, op, '1', ';']) else: tokens += JavaAndCPPProcessor.post_incre_decre_removal(code_string, post_expr, child, parser) return tokens @classmethod def extract_incre_decre_expression(cls, root, code_string): pre_expr = [] post_expr = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (((str(current_node.type) == '++') or (str(current_node.type) == '--')) and ((str(current_node.parent.type) == 'update_expression') or (str(current_node.parent.type) == 'postfix_unary_expression') or (str(current_node.parent.type) == 'prefix_unary_expression')) and (str(current_node.parent.parent.type) == 'assignment_expression')): nodes = current_node.parent.parent.children if ((len(nodes) == 3) and (str(nodes[0].type) == 'identifier')): if ((str(nodes[2].children[0].type) == '++') or (str(nodes[2].children[0].type) == '--')): pre_expr.append(current_node.parent.parent.parent) else: post_expr.append(current_node.parent.parent.parent) for child in current_node.children: queue.append(child) return (pre_expr, post_expr) @classmethod def handle_terminal_node(cls, root_node, code_string): if (root_node.type == 'comment'): str_const = '' else: str_const = code_string[root_node.start_byte:root_node.end_byte].decode('utf-8') return str_const @classmethod def remove_package_and_import(cls, code): if isinstance(code, str): code = code.encode() code = code.decode().split('\n') lines = [line.rstrip('\n') for line in code] current_code_lines = [] for line in lines: if (line.strip().startswith('import') or line.strip().startswith('package') or line.strip().startswith('#include')): continue current_code_lines.append(line) code = ('\n'.join(current_code_lines) if len(current_code_lines) else '') return code.encode() @classmethod def extract_expression(self, root, code): expressions = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'binary_expression'): children_nodes = current_node.children keep = ['<', '>', '<=', '>=', '==', '!='] counter = 0 for w in children_nodes: if (str(w.type) in keep): counter = (counter + 1) if (counter == 1): expressions.append(current_node) for child in current_node.children: queue.append(child) return expressions @classmethod def get_tokens_for_opswap(cls, code, root, left_oprd, operator, right_oprd): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == operator.start_byte) and (root.end_byte == operator.end_byte)): opt = code[operator.start_byte:operator.end_byte].decode() if (opt == '<'): tokens.append('>') elif (opt == '>'): tokens.append('<') elif (opt == '>='): tokens.append('<=') elif (opt == '<='): tokens.append('>=') elif (opt == '=='): tokens.append('==') elif (opt == '!='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: if ((child.start_byte == left_oprd.start_byte) and (child.end_byte == left_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, right_oprd, left_oprd, operator, right_oprd) elif ((child.start_byte == right_oprd.start_byte) and (child.end_byte == right_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, left_oprd, left_oprd, operator, right_oprd) else: (ts, _) = cls.get_tokens_for_opswap(code, child, left_oprd, operator, right_oprd) tokens += ts return (tokens, None) @classmethod def operand_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) expressions = cls.extract_expression(root, code) success = False try: while ((not success) and (len(expressions) > 0)): selected_exp = np.random.choice(expressions) expressions.remove(selected_exp) bin_exp = selected_exp condition = code[bin_exp.start_byte:bin_exp.end_byte].decode() bin_exp = bin_exp.children left_oprd = bin_exp[0] operator = bin_exp[1] right_oprd = bin_exp[2] try: code_list = cls.get_tokens_for_opswap(code, root, left_oprd, operator, right_oprd)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success) @classmethod def extract_if_else(cls, root, code_str, operator_list): ext_opt_list = ['&&', '&', '||', '|'] expressions = [] queue = [root] not_consider = [] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'if_statement'): clause = code_str[current_node.start_byte:current_node.end_byte].decode() des = current_node.children[1] cond = code_str[des.start_byte:des.end_byte].decode() stack = [des] nodes = [] while (len(stack) > 0): root1 = stack.pop() if (len(root1.children) == 0): nodes.append(root1) for child in root1.children: stack.append(child) nodes.reverse() counter = 0 extra_counter = 0 for w in nodes: if (str(w.type) in operator_list): counter = (counter + 1) if (str(w.type) in ext_opt_list): extra_counter = (extra_counter + 1) if (not ((counter == 1) and (extra_counter == 0))): continue children_nodes = current_node.children flagx = 0 flagy = 0 for w in children_nodes: if (str(w.type) == 'else'): flagx = 1 if (str(w.type) == 'if_statement'): not_consider.append(w) flagy = 1 if ((flagx == 1) and (flagy == 0)): expressions.append([current_node, des]) for child in current_node.children: if (child not in not_consider): queue.append(child) return expressions @classmethod def get_tokens_for_blockswap(cls, code, root, first_block, opt_node, second_block, flagx, flagy): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == opt_node.start_byte) and (root.end_byte == opt_node.end_byte)): op = code[root.start_byte:root.end_byte].decode() if (op == '<'): tokens.append('>=') elif (op == '>'): tokens.append('<=') elif (op == '>='): tokens.append('<') elif (op == '<='): tokens.append('>') elif (op == '!='): tokens.append('==') elif (op == '=='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if ((child.start_byte == first_block.start_byte) and (child.end_byte == first_block.end_byte) and (flagx == 0) and (str(child.type) == str(first_block.type))): flagx = 1 (ts, _) = cls.get_tokens_for_blockswap(code, second_block, first_block, opt_node, second_block, flagx, flagy) elif ((child.start_byte == second_block.start_byte) and (child.end_byte == second_block.end_byte) and (flagy == 0) and (str(child.type) == str(second_block.type))): flagy = 1 (ts, _) = cls.get_tokens_for_blockswap(code, first_block, first_block, opt_node, second_block, flagx, flagy) else: (ts, _) = cls.get_tokens_for_blockswap(code, child, first_block, opt_node, second_block, flagx, flagy) tokens += ts return (tokens, None) @classmethod def block_swap_java(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) operator_list = ['<', '>', '<=', '>=', '==', '!='] pair = cls.extract_if_else(root, code, operator_list) success = False lst = list(range(0, len(pair))) try: while ((not success) and (len(lst) > 0)): selected = np.random.choice(lst) lst.remove(selected) clause = pair[selected][0] des = pair[selected][1] st = [des] nodes = [] while (len(st) > 0): root1 = st.pop() if (len(root1.children) == 0): nodes.append(root1) if (code[root1.start_byte:root1.end_byte].decode() in operator_list): opt_node = root1 break for child in root1.children: st.append(child) nodes = clause.children flag = 0 for current_node in nodes: if (str(current_node.type) == 'block'): if (flag == 0): first_block = current_node flag = 1 else: second_block = current_node flagx = 0 flagy = 0 try: code_list = cls.get_tokens_for_blockswap(code, root, first_block, opt_node, second_block, flagx, flagy)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success) @classmethod def block_swap_c(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) operator_list = ['<', '>', '<=', '>=', '==', '!='] pair = cls.extract_if_else(root, code, operator_list) success = False lst = list(range(0, len(pair))) try: while ((not success) and (len(lst) > 0)): selected = np.random.choice(lst) lst.remove(selected) clause = pair[selected][0] des = pair[selected][1] st = [des] nodes = [] while (len(st) > 0): root1 = st.pop() if (len(root1.children) == 0): nodes.append(root1) if (code[root1.start_byte:root1.end_byte].decode() in operator_list): opt_node = root1 break for child in root1.children: st.append(child) nodes = clause.children flag = 0 for current_node in nodes: if (str(current_node.type) == 'compound_statement'): if (flag == 0): first_block = current_node flag = 1 else: second_block = current_node flagx = 0 flagy = 0 try: code_list = cls.get_tokens_for_blockswap(code, root, first_block, opt_node, second_block, flagx, flagy)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success)
class JavascriptProcessor(): @classmethod def create_dead_for_loop(cls, body): control_variable = ('_i_' + str(np.random.choice(list(range(10))))) p = np.random.uniform(0, 1) if (p < 0.5): prefix = (((((('for ( let ' + control_variable) + ' = 0 ; ') + control_variable) + ' > 0 ; ') + control_variable) + ' ++ ) { ') loop = ((prefix + body) + ' } ') return loop else: return (('for ( ; false ; ) { ' + body) + '}') @classmethod def create_dead_while_loop(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('while ( false ) { ' + body) + ' }') elif (p < 0.66): return (((((('while ( ' + control_variable) + ' < ') + control_variable) + ' ) { ') + body) + ' } ') else: return (((((('while ( ' + control_variable) + ' > ') + control_variable) + ' ) { ') + body) + ' } ') @classmethod def create_dead_if(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('if ( false ) { ' + body) + ' }') elif (p < 0.66): return (((((('if ( ' + control_variable) + ' < ') + control_variable) + ' ) { ') + body) + ' } ') else: return (((((('if ( ' + control_variable) + ' > ') + control_variable) + ' ) { ') + body) + ' } ') @classmethod def get_tokens_insert_before(cls, code_str, root, insertion_code, insert_before_node): if (not isinstance(insert_before_node, list)): insert_before_node = [insert_before_node] if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] if (root in insert_before_node): tokens += insertion_code.split() children = root.children if ((len(children) == 0) or (str(root.type) in ['string'])): tokens.append(code_str[root.start_byte:root.end_byte].decode()) else: for child in children: ts = cls.get_tokens_insert_before(code_str, child, insertion_code, insert_before_node) tokens += ts return tokens @classmethod def get_tokens(cls, code, root): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code[root.start_byte:root.end_byte].decode()] children = root.children if ((len(children) == 0) or (str(root.type) in ['string'])): tokens.append(code[root.start_byte:root.end_byte].decode()) else: for child in children: ts = cls.get_tokens(code, child) tokens += ts return tokens @classmethod def get_breaking_statements(cls, block): breakings = ['continue_statement', 'break_statement', 'return_statement'] statements = [] stack = [block] while (len(stack) > 0): top = stack.pop() if (str(top.type) in breakings): statements.append(top) else: for child in top.children: stack.append(child) return statements @classmethod def for_to_while_random(cls, code_string, parser): root = parser.parse_code(code_string) loops = cls.extract_for_loops(root) success = False try: while ((not success) and (len(loops) > 0)): selected_loop = np.random.choice(loops) loops.remove(selected_loop) (modified_root, modified_code_string, success) = cls.for_to_while(code_string, root, selected_loop, parser) if success: root = modified_root code_string = modified_code_string except: pass if (not success): ts = cls.get_tokens(code_string, root) code_string = cls.beautify_java_code(ts) return (root, code_string, success) @classmethod def while_to_for_random(cls, code_string, parser): root = parser.parse_code(code_string) loops = cls.extract_while_loops(root) success = False try: while ((not success) and (len(loops) > 0)): selected_loop = np.random.choice(loops) loops.remove(selected_loop) (modified_root, modified_code_string, success) = cls.while_to_for(code_string, root, selected_loop, parser) if success: root = modified_root code_string = modified_code_string if (not success): ts = cls.get_tokens(code_string, root) code_string = cls.beautify_java_code(ts) except: pass return (root, code_string, success) @classmethod def extract_for_loops(cls, root): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'for_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def beautify_java_code(cls, tokens): code = ' '.join(tokens) code = re.sub(' \\. ', '', code) code = re.sub(' \\+\\+', '++', code) return code @classmethod def get_tokens_replace_for(cls, code_str, for_node, root, init, cond, update, body): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): parent = root.parent return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if ((len(children) == 0) or (str(root.type) in ['string'])): tokens.append(code_str[root.start_byte:root.end_byte].decode()) else: for child in children: if (child == for_node): tokens.extend(((((((init + ['while', '(']) + cond) + [')', '{']) + body) + update) + ['}'])) else: tokens += cls.get_tokens_replace_for(code_str, for_node, child, init, cond, update, body) return tokens @classmethod def for_to_while(cls, code_string, root, fl, parser): children = fl.children init = children[2] init_tokens = cls.get_tokens(code_string, init) comparison = children[3] if (str(comparison.type) != ';'): comp_tokens = cls.get_tokens(code_string, comparison) if (comp_tokens[(- 1)] == ';'): comp_tokens = comp_tokens[:(- 1)] update = children[4] if (str(update.type) == ')'): update_tokens = [] body = children[5] else: update_tokens = (cls.get_tokens(code_string, update) + [';']) body = children[6] breaking_statements = cls.get_breaking_statements(body) body_tokens = cls.get_tokens_insert_before(code_string, body, ' '.join(update_tokens), breaking_statements) if ((len(body_tokens) >= 2) and ((body_tokens[0] == '{') and (body_tokens[(- 1)] == '}'))): body_tokens = body_tokens[1:(- 1)] tokens = cls.get_tokens_replace_for(code_str=code_string, for_node=fl, root=root, init=init_tokens, cond=comp_tokens, update=update_tokens, body=body_tokens) code = cls.beautify_java_code(tokens) return (parser.parse_code(code), code, True) return (root, code_string, False) @classmethod def extract_while_loops(cls, root): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'while_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def while_to_for(cls, code_string, root, wl, parser): children = wl.children condition = children[1] body = children[2] if (str(condition.type) == 'parenthesized_expression'): expr_tokens = cls.get_tokens(code_string, condition.children[1]) body_tokens = cls.get_tokens(code_string, body) if ((len(body_tokens) >= 2) and ((body_tokens[0] == '{') and (body_tokens[(- 1)] == '}'))): body_tokens = body_tokens[1:(- 1)] tokens = cls.get_tokens_replace_while(code_str=code_string, while_node=wl, root=root, cond=expr_tokens, body=body_tokens) code = cls.beautify_java_code(tokens) return (parser.parse_code(code), code, True) return (root, code_string, False) @classmethod def get_tokens_replace_while(cls, code_str, while_node, root, cond, body): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if ((len(children) == 0) or (str(root.type) in ['string'])): tokens.append(code_str[root.start_byte:root.end_byte].decode()) else: for child in children: if (child == while_node): tokens.extend(((((['for', '(', ';'] + cond) + [';', ')', '{']) + body) + ['}'])) else: tokens += cls.get_tokens_replace_while(code_str, while_node, child, cond, body) return tokens @classmethod def extract_expression(self, root, code): expressions = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'binary_expression'): children_nodes = current_node.children keep = ['<', '>', '<=', '>=', '==', '!=', '===', '!=='] counter = 0 for w in children_nodes: if (str(w.type) in keep): counter = (counter + 1) if (counter == 1): expressions.append(current_node) for child in current_node.children: queue.append(child) return expressions @classmethod def get_tokens_for_opswap(cls, code, root, left_oprd, operator, right_oprd): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == operator.start_byte) and (root.end_byte == operator.end_byte)): opt = code[operator.start_byte:operator.end_byte].decode() if (opt == '<'): tokens.append('>') elif (opt == '>'): tokens.append('<') elif (opt == '>='): tokens.append('<=') elif (opt == '<='): tokens.append('>=') elif (opt == '=='): tokens.append('==') elif (opt == '!='): tokens.append('!=') elif (opt == '==='): tokens.append('===') elif (opt == '!=='): tokens.append('!==') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: if ((child.start_byte == left_oprd.start_byte) and (child.end_byte == left_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, right_oprd, left_oprd, operator, right_oprd) elif ((child.start_byte == right_oprd.start_byte) and (child.end_byte == right_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, left_oprd, left_oprd, operator, right_oprd) else: (ts, _) = cls.get_tokens_for_opswap(code, child, left_oprd, operator, right_oprd) tokens += ts return (tokens, None) @classmethod def operand_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) expressions = cls.extract_expression(root, code) success = False try: while ((not success) and (len(expressions) > 0)): selected_exp = np.random.choice(expressions) expressions.remove(selected_exp) bin_exp = selected_exp condition = code[bin_exp.start_byte:bin_exp.end_byte].decode() bin_exp = bin_exp.children left_oprd = bin_exp[0] operator = bin_exp[1] right_oprd = bin_exp[2] try: code_list = cls.get_tokens_for_opswap(code, root, left_oprd, operator, right_oprd)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success) @classmethod def extract_if_else(cls, root, code_str, operator_list): ext_opt_list = ['&&', '&', '||', '|', 'and', 'or'] expressions = [] queue = [root] not_consider = [] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'if_statement'): clause = code_str[current_node.start_byte:current_node.end_byte].decode() des = current_node.children[1] cond = code_str[des.start_byte:des.end_byte].decode() stack = [des] nodes = [] while (len(stack) > 0): root1 = stack.pop() if (len(root1.children) == 0): nodes.append(root1) for child in root1.children: stack.append(child) nodes.reverse() counter = 0 extra_counter = 0 for w in nodes: if (str(w.type) in operator_list): counter = (counter + 1) if (str(w.type) in ext_opt_list): extra_counter = (extra_counter + 1) if (not ((counter == 1) and (extra_counter == 0))): continue children_nodes = current_node.children flagx = 0 flagy = 0 for w in children_nodes: if ((str(w.type) == 'else') or (str(w.type) == 'else_clause')): flagx = 1 m = w.children for x in m: if (str(x.type) == 'if_statement'): not_consider.append(x) flagy = 1 break if ((flagx == 1) and (flagy == 0)): expressions.append([current_node, des]) for child in current_node.children: if (child not in not_consider): queue.append(child) return expressions @classmethod def get_tokens_for_blockswap(cls, code, root, first_block, opt_node, second_block, flagx, flagy): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == opt_node.start_byte) and (root.end_byte == opt_node.end_byte)): op = code[root.start_byte:root.end_byte].decode() if (op == '<'): tokens.append('>=') elif (op == '>'): tokens.append('<=') elif (op == '>='): tokens.append('<') elif (op == '<='): tokens.append('>') elif (op == '!='): tokens.append('==') elif (op == '=='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if ((child.start_byte == first_block.start_byte) and (child.end_byte == first_block.end_byte) and (flagx == 0) and (str(child.type) == str(first_block.type))): flagx = 1 (ts, _) = cls.get_tokens_for_blockswap(code, second_block, first_block, opt_node, second_block, flagx, flagy) elif ((child.start_byte == second_block.start_byte) and (child.end_byte == second_block.end_byte) and (flagy == 0) and (str(child.type) == str(second_block.type))): flagy = 1 (ts, _) = cls.get_tokens_for_blockswap(code, first_block, first_block, opt_node, second_block, flagx, flagy) else: (ts, _) = cls.get_tokens_for_blockswap(code, child, first_block, opt_node, second_block, flagx, flagy) tokens += ts return (tokens, None) @classmethod def block_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) operator_list = ['<', '>', '<=', '>=', '==', '!='] pair = cls.extract_if_else(root, code, operator_list) success = False lst = list(range(0, len(pair))) try: while ((not success) and (len(lst) > 0)): selected = np.random.choice(lst) lst.remove(selected) clause = pair[selected][0] des = pair[selected][1] st = [des] nodes = [] while (len(st) > 0): root1 = st.pop() if (len(root1.children) == 0): nodes.append(root1) if (code[root1.start_byte:root1.end_byte].decode() in operator_list): opt_node = root1 break for child in root1.children: st.append(child) nodes = clause.children flag = 0 for current_node in nodes: if (str(current_node.type) == 'statement_block'): first_block = current_node elif (str(current_node.type) == 'else_clause'): new_list = current_node.children for w in new_list: if (str(w.type) == 'statement_block'): second_block = w break flagx = 0 flagy = 0 try: code_list = cls.get_tokens_for_blockswap(code, root, first_block, opt_node, second_block, flagx, flagy)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success)
class PhpProcessor(): @classmethod def create_dead_for_loop(cls, body): control_variable = ('$_i_' + str(np.random.choice(list(range(10))))) p = np.random.uniform(0, 1) if (p < 0.5): prefix = f'for ( {control_variable} = 0 ; {control_variable} > 0 ; {control_variable}++ )' loop = (((prefix + '{ ') + body) + '}') return loop else: return (('for ( ; FALSE ; ) { ' + body) + '}') @classmethod def create_dead_while_loop(cls, body): p = np.random.uniform(0, 1) control_variable = ('$_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('while ( FALSE ) { ' + body) + ' }') elif (p < 0.66): return (((((('while ( ' + control_variable) + ' < ') + control_variable) + ' ) { ') + body) + ' } ') else: return (((((('while ( ' + control_variable) + ' > ') + control_variable) + ' ) { ') + body) + ' } ') @classmethod def create_dead_if(cls, body): p = np.random.uniform(0, 1) control_variable = ('$_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('if ( FALSE ) { ' + body) + ' }') elif (p < 0.66): return (((((('if ( ' + control_variable) + ' < ') + control_variable) + ' ) { ') + body) + ' } ') else: return (((((('if ( ' + control_variable) + ' > ') + control_variable) + ' ) { ') + body) + ' } ') @classmethod def get_tokens_insert_before(cls, code_str, root, insertion_code, insert_before_node): if (not isinstance(insert_before_node, list)): insert_before_node = [insert_before_node] if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] if (root in insert_before_node): tokens += insertion_code.split() children = root.children if ((len(children) == 0) or (str(root.type) in ['variable_name', 'encapsed_string'])): tokens.append(code_str[root.start_byte:root.end_byte].decode()) else: for child in children: ts = cls.get_tokens_insert_before(code_str, child, insertion_code, insert_before_node) tokens += ts return tokens @classmethod def get_tokens(cls, code, root): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code[root.start_byte:root.end_byte].decode()] children = root.children if ((len(children) == 0) or (str(root.type) in ['variable_name', 'encapsed_string'])): tokens.append(code[root.start_byte:root.end_byte].decode()) else: for child in children: ts = cls.get_tokens(code, child) tokens += ts return tokens @classmethod def for_to_while_random(cls, code_string, parser): root = parser.parse_code(code_string) loops = cls.extract_for_loops(root) success = False try: while ((not success) and (len(loops) > 0)): selected_loop = np.random.choice(loops) loops.remove(selected_loop) (modified_root, modified_code_string, success) = cls.for_to_while(code_string, root, selected_loop, parser) if success: root = modified_root code_string = modified_code_string except: pass if (not success): ts = cls.get_tokens(code_string, root) code_string = cls.beautify_java_code(ts) return (root, code_string, success) @classmethod def while_to_for_random(cls, code_string, parser): root = parser.parse_code(code_string) loops = cls.extract_while_loops(root) success = False try: while ((not success) and (len(loops) > 0)): selected_loop = np.random.choice(loops) loops.remove(selected_loop) (modified_root, modified_code_string, success) = cls.while_to_for(code_string, root, selected_loop, parser) if success: root = modified_root code_string = modified_code_string if (not success): ts = cls.get_tokens(code_string, root) code_string = cls.beautify_java_code(ts) except: pass return (root, code_string, success) @classmethod def extract_for_loops(cls, root): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'for_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def beautify_java_code(cls, tokens): code = ' '.join(tokens) code = re.sub(' \\. ', '', code) code = re.sub(' \\+\\+', '++', code) return code @classmethod def get_tokens_replace_for(cls, code_str, for_node, root, init, cond, update, body): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if ((len(children) == 0) or (str(root.type) in ['variable_name', 'encapsed_string'])): tokens.append(code_str[root.start_byte:root.end_byte].decode()) else: for child in children: if (child == for_node): tokens.extend(((((((init + [';', 'while', '(']) + cond) + [')', '{']) + body) + update) + ['}'])) else: tokens += cls.get_tokens_replace_for(code_str, for_node, child, init, cond, update, body) return tokens @classmethod def get_breaking_statements(cls, block): breakings = ['continue_statement', 'break_statement', 'return_statement'] statements = [] stack = [block] while (len(stack) > 0): top = stack.pop() if (str(top.type) in breakings): statements.append(top) else: for child in top.children: stack.append(child) return statements @classmethod def for_to_while(cls, code_string, root, fl, parser): children = fl.children init = children[2] init_tokens = cls.get_tokens(code_string, init) comparison = children[4] if (str(comparison.type) != ';'): comp_tokens = cls.get_tokens(code_string, comparison) update = children[6] if (str(update.type) == ')'): update_tokens = [] body = children[7] else: update_tokens = (cls.get_tokens(code_string, update) + [';']) body = children[8] breaking_statements = cls.get_breaking_statements(body) body_tokens = cls.get_tokens_insert_before(code_string, body, ' '.join(update_tokens), breaking_statements) if ((len(body_tokens) >= 2) and ((body_tokens[0] == '{') and (body_tokens[(- 1)] == '}'))): body_tokens = body_tokens[1:(- 1)] tokens = cls.get_tokens_replace_for(code_str=code_string, for_node=fl, root=root, init=init_tokens, cond=comp_tokens, update=update_tokens, body=body_tokens) code = cls.beautify_java_code(tokens) return (parser.parse_code(code), code, True) return (root, code_string, False) @classmethod def extract_while_loops(cls, root): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'while_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def while_to_for(cls, code_string, root, wl, parser): children = wl.children condition = children[1] body = children[2] if (str(condition.type) == 'parenthesized_expression'): expr_tokens = cls.get_tokens(code_string, condition.children[1]) body_tokens = cls.get_tokens(code_string, body) if ((len(body_tokens) >= 2) and ((body_tokens[0] == '{') and (body_tokens[(- 1)] == '}'))): body_tokens = body_tokens[1:(- 1)] tokens = cls.get_tokens_replace_while(code_str=code_string, while_node=wl, root=root, cond=expr_tokens, body=body_tokens) code = cls.beautify_java_code(tokens) return (parser.parse_code(code), code, True) return (root, code_string, False) @classmethod def get_tokens_replace_while(cls, code_str, while_node, root, cond, body): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if ((len(children) == 0) or (str(root.type) in ['variable_name', 'encapsed_string'])): tokens.append(code_str[root.start_byte:root.end_byte].decode()) else: for child in children: if (child == while_node): tokens.extend(((((['for', '(', ';'] + cond) + [';', ')', '{']) + body) + ['}'])) else: tokens += cls.get_tokens_replace_while(code_str, while_node, child, cond, body) return tokens @classmethod def extract_expression(self, root, code): expressions = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'binary_expression'): children_nodes = current_node.children keep = ['<', '>', '<=', '>=', '==', '!=', '<>', '===', '!=='] counter = 0 for w in children_nodes: if (str(w.type) in keep): counter = (counter + 1) if (counter == 1): expressions.append(current_node) for child in current_node.children: queue.append(child) return expressions @classmethod def get_tokens_for_opswap(cls, code, root, left_oprd, operator, right_oprd): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == operator.start_byte) and (root.end_byte == operator.end_byte)): opt = code[operator.start_byte:operator.end_byte].decode() if (opt == '<'): tokens.append('>') elif (opt == '>'): tokens.append('<') elif (opt == '>='): tokens.append('<=') elif (opt == '<='): tokens.append('>=') elif (opt == '=='): tokens.append('==') elif (opt == '!='): tokens.append('!=') elif (opt == '<>'): tokens.append('<>') elif (opt == '==='): tokens.append('===') elif (opt == '!=='): tokens.append('!==') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: if ((child.start_byte == left_oprd.start_byte) and (child.end_byte == left_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, right_oprd, left_oprd, operator, right_oprd) elif ((child.start_byte == right_oprd.start_byte) and (child.end_byte == right_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, left_oprd, left_oprd, operator, right_oprd) else: (ts, _) = cls.get_tokens_for_opswap(code, child, left_oprd, operator, right_oprd) tokens += ts return (tokens, None) @classmethod def operand_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) expressions = cls.extract_expression(root, code) success = False try: while ((not success) and (len(expressions) > 0)): selected_exp = np.random.choice(expressions) expressions.remove(selected_exp) bin_exp = selected_exp condition = code[bin_exp.start_byte:bin_exp.end_byte].decode() bin_exp = bin_exp.children left_oprd = bin_exp[0] operator = bin_exp[1] right_oprd = bin_exp[2] try: code_list = cls.get_tokens_for_opswap(code, root, left_oprd, operator, right_oprd)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success) @classmethod def extract_if_else(cls, root, code_str, operator_list): ext_opt_list = ['&&', '&', '||', '|'] expressions = [] queue = [root] not_consider = [] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'if_statement'): clause = code_str[current_node.start_byte:current_node.end_byte].decode() des = current_node.children[1] cond = code_str[des.start_byte:des.end_byte].decode() stack = [des] nodes = [] while (len(stack) > 0): root1 = stack.pop() if (len(root1.children) == 0): nodes.append(root1) for child in root1.children: stack.append(child) nodes.reverse() counter = 0 extra_counter = 0 for w in nodes: if (str(w.type) in operator_list): counter = (counter + 1) if (str(w.type) in ext_opt_list): extra_counter = (extra_counter + 1) if (not ((counter == 1) and (extra_counter == 0))): continue children_nodes = current_node.children flagx = 0 flagy = 0 for w in children_nodes: if (str(w.type) == 'else_clause'): flagx = 1 if (str(w.type) == 'else_if_clause'): not_consider.append(w) flagy = 1 if ((flagx == 1) and (flagy == 0)): expressions.append([current_node, des]) for child in current_node.children: if (child not in not_consider): queue.append(child) return expressions @classmethod def get_tokens_for_blockswap(cls, code, root, first_block, opt_node, second_block, flagx, flagy): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == opt_node.start_byte) and (root.end_byte == opt_node.end_byte)): op = code[root.start_byte:root.end_byte].decode() if (op == '<'): tokens.append('>=') elif (op == '>'): tokens.append('<=') elif (op == '>='): tokens.append('<') elif (op == '<='): tokens.append('>') elif (op == '!='): tokens.append('==') elif (op == '=='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if ((child.start_byte == first_block.start_byte) and (child.end_byte == first_block.end_byte) and (flagx == 0) and (str(child.type) == str(first_block.type))): flagx = 1 (ts, _) = cls.get_tokens_for_blockswap(code, second_block, first_block, opt_node, second_block, flagx, flagy) elif ((child.start_byte == second_block.start_byte) and (child.end_byte == second_block.end_byte) and (flagy == 0) and (str(child.type) == str(second_block.type))): flagy = 1 (ts, _) = cls.get_tokens_for_blockswap(code, first_block, first_block, opt_node, second_block, flagx, flagy) else: (ts, _) = cls.get_tokens_for_blockswap(code, child, first_block, opt_node, second_block, flagx, flagy) tokens += ts return (tokens, None) @classmethod def block_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) operator_list = ['<', '>', '<=', '>=', '==', '!='] pair = cls.extract_if_else(root, code, operator_list) success = False lst = list(range(0, len(pair))) try: while ((not success) and (len(lst) > 0)): selected = np.random.choice(lst) lst.remove(selected) clause = pair[selected][0] des = pair[selected][1] st = [des] nodes = [] while (len(st) > 0): root1 = st.pop() if (len(root1.children) == 0): nodes.append(root1) if (code[root1.start_byte:root1.end_byte].decode() in operator_list): opt_node = root1 break for child in root1.children: st.append(child) nodes = clause.children flag = 0 for current_node in nodes: if (str(current_node.type) == 'compound_statement'): first_block = current_node elif (str(current_node.type) == 'else_clause'): second_block = current_node.children[1] flagx = 0 flagy = 0 try: code_list = cls.get_tokens_for_blockswap(code, root, first_block, opt_node, second_block, flagx, flagy)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success)
class PythonProcessor(): @classmethod def create_dead_for_loop(cls, body): control_variable = ('_i_' + str(np.random.choice(list(range(10))))) loop = f'NEWLINE for {control_variable} in range ( 0 ) : NEWLINE INDENT {body} NEWLINE DEDENT ' return loop @classmethod def create_dead_while_loop(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return f'while False : NEWLINE INDENT {body} NEWLINE DEDENT' elif (p < 0.66): return f'while {control_variable} < {control_variable} : NEWLINE INDENT {body} NEWLINE DEDENT' else: return f'while {control_variable} > {control_variable} : NEWLINE INDENT {body} NEWLINE DEDENT' @classmethod def create_dead_if(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return f'if False : NEWLINE INDENT {body} NEWLINE DEDENT' elif (p < 0.66): return f'if {control_variable} < {control_variable} : NEWLINE INDENT {body} NEWLINE DEDENT' else: return f'if {control_variable} > {control_variable} : NEWLINE INDENT {body} NEWLINE DEDENT' @classmethod def get_tokens_insert_before(cls, code_str, root, insertion_code, insert_before_node): if (not isinstance(insert_before_node, list)): insert_before_node = [insert_before_node] if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): parent = root.parent if (len(parent.children) == 1): return tokens else: return [code_str[root.start_byte:root.end_byte].decode()] if (root in insert_before_node): tokens += insertion_code.split() children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if (child_type == 'block'): tokens += ['NEWLINE', 'INDENT'] ts = cls.get_tokens_insert_before(code_str, child, insertion_code, insert_before_node) tokens += ts if child_type.endswith('statement'): tokens.append('NEWLINE') elif (child_type == 'block'): tokens.append('DEDENT') return tokens @classmethod def get_tokens(cls, code, root): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): parent = root.parent if (len(parent.children) == 1): return tokens else: return [code[root.start_byte:root.end_byte].decode()] children = root.children if (len(children) == 0): tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if (child_type == 'block'): tokens += ['NEWLINE', 'INDENT'] ts = cls.get_tokens(code, child) tokens += ts if child_type.endswith('statement'): tokens.append('NEWLINE') elif (child_type == 'block'): tokens.append('DEDENT') return tokens @classmethod def for_to_while_random(cls, code_string, parser): root = parser.parse_code(code_string) loops = cls.extract_for_loops(root, code_string) success = False try: while ((not success) and (len(loops) > 0)): selected_loop = np.random.choice(loops) loops.remove(selected_loop) (modified_root, modified_code_string, success) = cls.for_to_while(code_string, root, selected_loop, parser) if success: root = modified_root code_string = modified_code_string except: pass if (not success): code_string = cls.beautify_python_code(cls.get_tokens(code_string, root)) else: code_string = cls.beautify_python_code(code_string.split()) return (root, code_string, success) @classmethod def while_to_for_random(cls, code_string, parser): root = parser.parse_code(code_string) return (root, code_string, False) @classmethod def extract_for_loops(cls, root, code_str): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'for_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def beautify_python_code(cls, tokens): indent_count = 0 code = '' i = 0 while (i < len(tokens)): token = tokens[i] if (token == 'NEWLINE'): code += '\n' for _ in range(indent_count): code += '\t' elif (token == 'INDENT'): indent_count += 1 code += '\t' elif (token == 'DEDENT'): indent_count -= 1 if (code[(- 1)] == '\t'): code = code[:(- 1)] else: code += (token + ' ') i += 1 lines = code.split('\n') taken_lines = [] for line in lines: if (len(line.strip()) > 0): taken_lines.append(line.rstrip()) code = '\n'.join(taken_lines) return code @classmethod def get_tokens_replace_for(cls, code_str, for_node, root, while_node): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): parent = root.parent if (len(parent.children) == 1): return tokens else: return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode()) for child in children: if (child == for_node): tokens += while_node else: child_type = str(child.type) if (child_type == 'block'): tokens += ['NEWLINE', 'INDENT'] tokens += cls.get_tokens_replace_for(code_str, for_node, child, while_node) if child_type.endswith('statement'): tokens.append('NEWLINE') elif (child_type == 'block'): tokens.append('DEDENT') return tokens @classmethod def for_to_while(cls, code_string, root, fl, parser): try: identifier = fl.children[1] in_node = fl.children[2] range_node = fl.children[3] body_node = fl.children[5] range_function = range_node.children[0] range_function_name = cls.get_tokens(code_string, range_function)[0] if ((range_function_name == 'range') and ((str(identifier.type) == 'identifier') and (len(identifier.children) == 0)) and ((str(in_node.type) == 'in') and (len(in_node.children) == 0))): argument_list = range_node.children[1].children args = [] for a in argument_list: k = str(a.type) if (k not in ['(', ',', ')']): args.append(a) (start, stop, step) = (['0'], ['0'], ['1']) if (len(args) == 1): stop = cls.get_tokens(code_string, args[0]) elif (len(args) == 2): start = cls.get_tokens(code_string, args[0]) stop = cls.get_tokens(code_string, args[1]) else: start = cls.get_tokens(code_string, args[0]) stop = cls.get_tokens(code_string, args[1]) step = cls.get_tokens(code_string, args[2]) identifier_name = cls.get_tokens(code_string, identifier)[0] terminal_statements = cls.find_terminal_statement(body_node) body_tokens = cls.get_tokens_insert_before(code_string, body_node, ' '.join((([identifier_name, '+='] + step) + ['NEWLINE'])), terminal_statements) while_stmt = ((((((((([identifier_name, '='] + start) + ['NEWLINE']) + ['while', identifier_name, 'in', 'list', '(', 'range', '(']) + stop) + [')', ')', ':', 'NEWLINE', 'INDENT']) + body_tokens) + ['NEWLINE', identifier_name, '+=']) + step) + ['DEDENT', 'NEWLINE']) tokens = cls.get_tokens_replace_for(code_str=code_string, for_node=fl, while_node=while_stmt, root=root) code = cls.beautify_python_code(tokens) return (parser.parse_code(code), ' '.join(tokens), True) except: pass return (root, code_string, False) @classmethod def find_terminal_statement(cls, body_node): statements = ['continue_statement', 'break_statement', 'return_statement'] terminals = [] stack = [body_node] while (len(stack) > 0): top = stack.pop() if (str(top.type) in statements): terminals.append(top) else: for child in top.children: stack.append(child) return terminals pass @classmethod def extract_while_loops(cls, root): loops = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'while_statement'): loops.append(current_node) for child in current_node.children: queue.append(child) return loops @classmethod def while_to_for(cls, code_string, root, wl, parser): return (root, code_string, False) @classmethod def get_tokens_replace_while(cls, code_str, while_node, root, cond, body): raise NotImplementedError @classmethod def extract_expression(self, root, code): expressions = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'comparison_operator'): children_nodes = current_node.children keep = ['<', '>', '<=', '>=', '==', '!='] counter = 0 for w in children_nodes: if (str(w.type) in keep): counter = (counter + 1) if (counter == 1): expressions.append(current_node) for child in current_node.children: queue.append(child) return expressions @classmethod def get_tokens_for_opswap(cls, code, root, left_oprd, operator, right_oprd): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): parent = root.parent if (len(parent.children) == 1): return (tokens, None) else: return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == operator.start_byte) and (root.end_byte == operator.end_byte)): opt = code[operator.start_byte:operator.end_byte].decode() if (opt == '<'): tokens.append('>') elif (opt == '>'): tokens.append('<') elif (opt == '>='): tokens.append('<=') elif (opt == '<='): tokens.append('>=') elif (opt == '=='): tokens.append('==') elif (opt == '!='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if (child_type == 'block'): tokens += ['NEWLINE', 'INDENT'] if ((child.start_byte == left_oprd.start_byte) and (child.end_byte == left_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, right_oprd, left_oprd, operator, right_oprd) elif ((child.start_byte == right_oprd.start_byte) and (child.end_byte == right_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, left_oprd, left_oprd, operator, right_oprd) else: (ts, _) = cls.get_tokens_for_opswap(code, child, left_oprd, operator, right_oprd) tokens += ts if child_type.endswith('statement'): tokens.append('NEWLINE') elif (child_type == 'block'): tokens.append('DEDENT') return (tokens, None) @classmethod def operand_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) expressions = cls.extract_expression(root, code) success = False try: while ((not success) and (len(expressions) > 0)): selected_exp = np.random.choice(expressions) expressions.remove(selected_exp) bin_exp = selected_exp condition = code[bin_exp.start_byte:bin_exp.end_byte].decode() bin_exp = bin_exp.children left_oprd = bin_exp[0] operator = bin_exp[1] right_oprd = bin_exp[2] try: code_list = cls.get_tokens_for_opswap(code, root, left_oprd, operator, right_oprd)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_python_code(cls.get_tokens(code_str, root)) else: code_string = cls.beautify_python_code(code_string.split()) return (code_string, success) @classmethod def extract_if_else(cls, root, code_str, operator_list): ext_opt_list = ['&&', '&', '||', '|'] expressions = [] queue = [root] not_consider = [] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'if_statement'): clause = code_str[current_node.start_byte:current_node.end_byte].decode() des = current_node.children[1] cond = code_str[des.start_byte:des.end_byte].decode() stack = [des] nodes = [] while (len(stack) > 0): root1 = stack.pop() if (len(root1.children) == 0): nodes.append(root1) for child in root1.children: stack.append(child) nodes.reverse() counter = 0 extra_counter = 0 for w in nodes: if (str(w.type) in operator_list): counter = (counter + 1) if (str(w.type) in ext_opt_list): extra_counter = (extra_counter + 1) if (not ((counter == 1) and (extra_counter == 0))): continue children_nodes = current_node.children flagx = 0 flagy = 0 for w in children_nodes: if (str(w.type) == 'else_clause'): flagx = 1 if (str(w.type) == 'elif_clause'): flagy = 1 if ((flagx == 1) and (flagy == 0)): expressions.append([current_node, des]) for child in current_node.children: if (child not in not_consider): queue.append(child) return expressions @classmethod def get_tokens_for_blockswap(cls, code, root, first_block, opt_node, second_block, flagx, flagy): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): parent = root.parent if (len(parent.children) == 1): return (tokens, None) else: return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == opt_node.start_byte) and (root.end_byte == opt_node.end_byte)): op = code[root.start_byte:root.end_byte].decode() if (op == '<'): tokens.append('>=') elif (op == '>'): tokens.append('<=') elif (op == '>='): tokens.append('<') elif (op == '<='): tokens.append('>') elif (op == '!='): tokens.append('==') elif (op == '=='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if (child_type == 'block'): tokens += ['NEWLINE', 'INDENT'] if ((child.start_byte == first_block.start_byte) and (child.end_byte == first_block.end_byte) and (flagx == 0) and (str(child.type) == str(first_block.type))): flagx = 1 (ts, _) = cls.get_tokens_for_blockswap(code, second_block, first_block, opt_node, second_block, flagx, flagy) elif ((child.start_byte == second_block.start_byte) and (child.end_byte == second_block.end_byte) and (flagy == 0) and (str(child.type) == str(second_block.type))): flagy = 1 (ts, _) = cls.get_tokens_for_blockswap(code, first_block, first_block, opt_node, second_block, flagx, flagy) else: (ts, _) = cls.get_tokens_for_blockswap(code, child, first_block, opt_node, second_block, flagx, flagy) tokens += ts if child_type.endswith('statement'): tokens.append('NEWLINE') elif (child_type == 'block'): tokens.append('DEDENT') return (tokens, None) @classmethod def block_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) operator_list = ['<', '>', '<=', '>=', '==', '!='] pair = cls.extract_if_else(root, code, operator_list) success = False lst = list(range(0, len(pair))) try: while ((not success) and (len(lst) > 0)): selected = np.random.choice(lst) lst.remove(selected) clause = pair[selected][0] des = pair[selected][1] st = [des] nodes = [] while (len(st) > 0): root1 = st.pop() if (len(root1.children) == 0): nodes.append(root1) if (code[root1.start_byte:root1.end_byte].decode() in operator_list): opt_node = root1 break for child in root1.children: st.append(child) nodes = clause.children flag = 0 for current_node in nodes: if (str(current_node.type) == 'block'): first_block = current_node elif (str(current_node.type) == 'else_clause'): new_list = current_node.children for w in new_list: if (str(w.type) == 'block'): second_block = w break flagx = 0 flagy = 0 try: code_list = cls.get_tokens_for_blockswap(code, root, first_block, opt_node, second_block, flagx, flagy)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_python_code(cls.get_tokens(code_str, root)) else: code_string = cls.beautify_python_code(code_string.split()) return (code_string, success)
def get_python_tokens(code, root=None): if isinstance(code, bytes): code = code.decode() tokens = [] for token in tokenize.tokenize(BytesIO(code.encode('utf-8')).readline): if ((token.type == 0) or (token.type >= 58)): continue elif (token.type == 4): tokens.append('NEWLINE') elif (token.type == 5): tokens.append('INDENT') elif (token.type == 6): tokens.append('DEDENT') else: tokens.append(token.string) return (tokens, None)
class RubyProcessor(): @classmethod def create_dead_for_loop(cls, body): control_variable = ('_i_' + str(np.random.choice(list(range(10))))) return f'for {control_variable} in 0..0 do {body} end ' @classmethod def create_dead_while_loop(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('until false do ' + body) + ' end ') elif (p < 0.66): return (((((('until ' + control_variable) + ' < ') + control_variable) + ' do ') + body) + ' end ') else: return (((((('until ' + control_variable) + ' > ') + control_variable) + ' do ') + body) + ' end ') @classmethod def create_dead_if(cls, body): p = np.random.uniform(0, 1) control_variable = ('_i_' + str(np.random.choice(list(range(10))))) if (p < 0.33): return (('if false then ' + body) + ' end ') elif (p < 0.66): return (((((('if ' + control_variable) + ' < ') + control_variable) + ' then ') + body) + ' end ') else: return (((((('if ' + control_variable) + ' > ') + control_variable) + ' then ') + body) + ' end ') @classmethod def extract_expression(self, root, code): expressions = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (str(current_node.type) == 'binary'): children_nodes = current_node.children keep = ['<', '>', '<=', '>=', '==', '!=', '==='] counter = 0 for w in children_nodes: if (str(w.type) in keep): counter = (counter + 1) if (counter == 1): expressions.append(current_node) for child in current_node.children: queue.append(child) return expressions @classmethod def get_tokens_for_opswap(cls, code, root, left_oprd, operator, right_oprd): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == operator.start_byte) and (root.end_byte == operator.end_byte)): opt = code[operator.start_byte:operator.end_byte].decode() if (opt == '<'): tokens.append('>') elif (opt == '>'): tokens.append('<') elif (opt == '>='): tokens.append('<=') elif (opt == '<='): tokens.append('>=') elif (opt == '=='): tokens.append('==') elif (opt == '!='): tokens.append('!=') elif (opt == '==='): tokens.append('===') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: if ((child.start_byte == left_oprd.start_byte) and (child.end_byte == left_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, right_oprd, left_oprd, operator, right_oprd) elif ((child.start_byte == right_oprd.start_byte) and (child.end_byte == right_oprd.end_byte)): (ts, _) = cls.get_tokens_for_opswap(code, left_oprd, left_oprd, operator, right_oprd) else: (ts, _) = cls.get_tokens_for_opswap(code, child, left_oprd, operator, right_oprd) tokens += ts return (tokens, None) @classmethod def operand_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) expressions = cls.extract_expression(root, code) success = False try: while ((not success) and (len(expressions) > 0)): selected_exp = np.random.choice(expressions) expressions.remove(selected_exp) bin_exp = selected_exp condition = code[bin_exp.start_byte:bin_exp.end_byte].decode() bin_exp = bin_exp.children left_oprd = bin_exp[0] operator = bin_exp[1] right_oprd = bin_exp[2] try: code_list = cls.get_tokens_for_opswap(code, root, left_oprd, operator, right_oprd)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = get_tokens(code_str, root) return (code_string, success) @classmethod def extract_if_else(cls, root, code_str, operator_list): ext_opt_list = ['&&', '&', '||', '|', 'and', 'or'] expressions = [] queue = [root] not_consider = [] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (((str(current_node.type) == 'if') or (str(current_node.type) == 'unless')) and (len(code_str[current_node.start_byte:current_node.end_byte].decode()) > 6)): clause = code_str[current_node.start_byte:current_node.end_byte].decode() des = current_node.children[1] cond = code_str[des.start_byte:des.end_byte].decode() stack = [des] nodes = [] while (len(stack) > 0): root1 = stack.pop() if (len(root1.children) == 0): nodes.append(root1) for child in root1.children: stack.append(child) nodes.reverse() counter = 0 extra_counter = 0 for w in nodes: if (str(w.type) in operator_list): counter = (counter + 1) if (str(w.type) in ext_opt_list): extra_counter = (extra_counter + 1) if ((counter > 1) or (extra_counter > 0)): continue children_nodes = current_node.children flagx = 0 flagy = 0 for w in children_nodes: if (str(w.type) == 'else'): flagx = 1 if (str(w.type) == 'elsif'): not_consider.append(w) flagy = 1 if ((flagx == 1) and (flagy == 0)): expressions.append([current_node, des]) for child in current_node.children: if (child not in not_consider): queue.append(child) return expressions @classmethod def get_tokens_for_blockswap(cls, code, root, first_block, opt_node, second_block, flagx, flagy): if isinstance(code, str): code = code.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return (tokens, None) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], None) children = root.children if (len(children) == 0): if ((root.start_byte == opt_node.start_byte) and (root.end_byte == opt_node.end_byte)): op = code[root.start_byte:root.end_byte].decode() if (op == '<'): tokens.append('>=') elif (op == '>'): tokens.append('<=') elif (op == '>='): tokens.append('<') elif (op == '<='): tokens.append('>') elif (op == '!='): tokens.append('==') elif (op == '=='): tokens.append('!=') else: tokens.append(code[root.start_byte:root.end_byte].decode()) for child in children: child_type = str(child.type) if ((child.start_byte == first_block.start_byte) and (child.end_byte == first_block.end_byte) and (flagx == 0) and (str(child.type) == str(first_block.type))): flagx = 1 (ts, _) = cls.get_tokens_for_blockswap(code, second_block, first_block, opt_node, second_block, flagx, flagy) elif ((child.start_byte == second_block.start_byte) and (child.end_byte == second_block.end_byte) and (flagy == 0) and (str(child.type) == str(second_block.type))): flagy = 1 (ts, _) = cls.get_tokens_for_blockswap(code, first_block, first_block, opt_node, second_block, flagx, flagy) else: (ts, _) = cls.get_tokens_for_blockswap(code, child, first_block, opt_node, second_block, flagx, flagy) tokens += ts return (tokens, None) @classmethod def block_swap(cls, code_str, parser): code = code_str.encode() root = parser.parse_code(code) operator_list = ['<', '>', '<=', '>=', '==', '!='] pair = cls.extract_if_else(root, code, operator_list) success = False lst = list(range(0, len(pair))) try: while ((not success) and (len(lst) > 0)): selected = np.random.choice(lst) lst.remove(selected) clause = pair[selected][0] des = pair[selected][1] st = [des] nodes = [] while (len(st) > 0): root1 = st.pop() if (len(root1.children) == 0): nodes.append(root1) if (code[root1.start_byte:root1.end_byte].decode() in operator_list): opt_node = root1 break for child in root1.children: st.append(child) nodes = clause.children flag = 0 for current_node in nodes: if (str(current_node.type) == 'then'): first_block = current_node elif (str(current_node.type) == 'else'): second_block = current_node.children[1] flagx = 0 flagy = 0 try: code_list = cls.get_tokens_for_blockswap(code, root, first_block, opt_node, second_block, flagx, flagy)[0] code_string = '' for w in code_list: code_string = ((code_string + w) + ' ') code_string = code_string.strip() success = True except: success = False continue except: pass if (not success): code_string = cls.beautify_java_code(get_tokens(code_str, root)) return (code_string, success) @classmethod def beautify_java_code(cls, tokens): code = ' '.join(tokens) code = re.sub(' \\. ', '', code) code = re.sub(' \\+\\+', '++', code) return code
def get_tokens(code_str, root): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode().strip()) for child in children: tokens += get_tokens(code_str, child) return tokens
def get_tokens_insert_before(code_str, root, insertion_code, insert_before_node): if isinstance(code_str, str): code_str = code_str.encode() assert isinstance(root, Node) tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): return [code_str[root.start_byte:root.end_byte].decode()] if (root == insert_before_node): tokens += insertion_code.split() children = root.children if (len(children) == 0): tokens.append(code_str[root.start_byte:root.end_byte].decode()) for child in children: tokens += get_tokens_insert_before(code_str, child, insertion_code, insert_before_node) return tokens
def dfs_print(root, level=0): for _ in range(level): print('\t', end='') print(root) for child in root.children: dfs_print(child, (level + 1))
def count_nodes(root): num_nodes = 1 for child in root.children: if (child is not None): num_nodes += count_nodes(child) return num_nodes
def extract_statement_within_size(root, max_node=10, endswith=None, code_string=None, tokenizer=None): if (endswith is None): endswith = ['statement'] statements = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] node_count = count_nodes(current_node) if ((code_string is not None) and (tokenizer is not None)): tokens = tokenizer(code_string, current_node) current_code = ' '.join(tokens).strip() else: current_code = 'please provide code string and tokenizer to analyze code length' if (any((str(current_node.type).endswith(e) for e in endswith)) and (1 < node_count < max_node) and (len(current_code) > 0)): statements.append(current_node) for child in current_node.children: queue.append(child) return statements
class BlockSwap(TransformationBase): '\n Swapping if_else block\n ' def __init__(self, parser_path, language): super(BlockSwap, self).__init__(parser_path=parser_path, language=language) self.language = language self.transformations = processor_function[language] processor_map = {'java': self.get_tokens_with_node_type, 'c': self.get_tokens_with_node_type, 'cpp': self.get_tokens_with_node_type, 'c_sharp': self.get_tokens_with_node_type, 'javascript': JavascriptProcessor.get_tokens, 'python': PythonProcessor.get_tokens, 'php': PhpProcessor.get_tokens, 'ruby': self.get_tokens_with_node_type, 'go': self.get_tokens_with_node_type} self.final_processor = processor_map[self.language] def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: success = False transform_functions = copy.deepcopy(self.transformations) while ((not success) and (len(transform_functions) > 0)): function = np.random.choice(transform_functions) transform_functions.remove(function) (modified_code, success) = function(code, self) if success: code = modified_code root_node = self.parse_code(code=code) return_values = self.final_processor(code=code.encode(), root=root_node) if isinstance(return_values, tuple): (tokens, types) = return_values else: (tokens, types) = (return_values, None) return (re.sub('[ \t\n]+', ' ', ' '.join(tokens)), {'types': types, 'success': success})
class ConfusionRemover(TransformationBase): '\n Change the `for` loops with `while` loops and vice versa.\n ' def __init__(self, parser_path, language): super(ConfusionRemover, self).__init__(parser_path=parser_path, language=language) self.language = language if (language in processor_function): self.transformations = processor_function[language] else: self.transformations = [] processor_map = {'java': self.get_tokens_with_node_type, 'c': self.get_tokens_with_node_type, 'cpp': self.get_tokens_with_node_type, 'c_sharp': self.get_tokens_with_node_type, 'javascript': JavascriptProcessor.get_tokens, 'python': PythonProcessor.get_tokens, 'php': PhpProcessor.get_tokens, 'ruby': self.get_tokens_with_node_type, 'go': self.get_tokens_with_node_type} self.final_processor = processor_map[self.language] def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: success = False transform_functions = copy.deepcopy(self.transformations) while ((not success) and (len(transform_functions) > 0)): function = np.random.choice(transform_functions) transform_functions.remove(function) (modified_root, modified_code, success) = function(code, self) if success: code = modified_code root_node = self.parse_code(code=code) return_values = self.final_processor(code=code.encode(), root=root_node) if isinstance(return_values, tuple): (tokens, types) = return_values else: (tokens, types) = (return_values, None) return (' '.join(tokens), {'types': types, 'success': success})
class DeadCodeInserter(TransformationBase): def __init__(self, parser_path: str, language: str): super(DeadCodeInserter, self).__init__(parser_path=parser_path, language=language) self.language = language self.processor = processor_function[self.language] self.tokenizer_function = tokenizer_function[self.language] self.insertion_function = insertion_function[self.language] def insert_random_dead_code(self, code_string, max_node_in_statement=(- 1)): root = self.parse_code(code_string) original_node_count = count_nodes(root) if (max_node_in_statement == (- 1)): max_node_in_statement = int((original_node_count / 2)) if (self.language == 'ruby'): statement_markers = ['assignment', 'until', 'call', 'if', 'for', 'while'] else: statement_markers = None statements = extract_statement_within_size(root, max_node_in_statement, statement_markers, code_string=code_string, tokenizer=self.tokenizer_function) original_code = ' '.join(self.tokenizer_function(code_string, root)) try: while (len(statements) > 0): (random_stmt, insert_before) = np.random.choice(statements, 2) statements.remove(random_stmt) dead_coed_body = ' '.join(self.tokenizer_function(code_string, random_stmt)).strip() dead_code_function = np.random.choice([self.processor.create_dead_for_loop, self.processor.create_dead_while_loop, self.processor.create_dead_if]) dead_code = dead_code_function(dead_coed_body) modified_code = ' '.join(self.insertion_function(code_str=code_string, root=root, insertion_code=dead_code, insert_before_node=insert_before)) if (modified_code != original_code): modified_root = self.parse_code(' '.join(modified_code)) return (modified_root, modified_code, True) except: pass return (root, original_code, False) def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: (root, code, success) = self.insert_random_dead_code(code, (- 1)) code = re.sub('[ \n\t]+', ' ', code) return (code, {'success': success})
class DemoTransformation(TransformationBase): def __init__(self, parser, language): super(DemoTransformation, self).__init__(parser_path=parser, language=language) def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: root_node = self.parse_code(code=code) (tokens, types) = self.get_tokens_with_node_type(code=code.encode(), root=root_node) return (' '.join(tokens), types)
class ForWhileTransformer(TransformationBase): '\n Change the `for` loops with `while` loops and vice versa.\n ' def __init__(self, parser_path, language): super(ForWhileTransformer, self).__init__(parser_path=parser_path, language=language) self.language = language self.transformations = processor_function[language] processor_map = {'java': self.get_tokens_with_node_type, 'c': self.get_tokens_with_node_type, 'cpp': self.get_tokens_with_node_type, 'c_sharp': self.get_tokens_with_node_type, 'javascript': JavascriptProcessor.get_tokens, 'python': PythonProcessor.get_tokens, 'php': PhpProcessor.get_tokens, 'ruby': self.get_tokens_with_node_type, 'go': self.get_tokens_with_node_type} self.final_processor = processor_map[self.language] def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: success = False transform_functions = copy.deepcopy(self.transformations) while ((not success) and (len(transform_functions) > 0)): function = np.random.choice(transform_functions) transform_functions.remove(function) (modified_root, modified_code, success) = function(code, self) if success: code = modified_code root_node = self.parse_code(code=code) return_values = self.final_processor(code=code.encode(), root=root_node) if isinstance(return_values, tuple): (tokens, types) = return_values else: (tokens, types) = (return_values, None) return (re.sub('[ \t\n]+', ' ', ' '.join(tokens)), {'types': types, 'success': success})
class NoTransformation(TransformationBase): def __init__(self, parser_path: str, language: str) -> object: super().__init__(parser_path, language) if (not os.path.exists(parser_path)): raise ValueError(f'Language parser does not exist at {parser_path}. Please run `setup.sh` to properly set the environment!') self.lang_object = Language(parser_path, language) self.parser = Parser() self.parser.set_language(self.lang_object) processor_map = {'java': self.get_tokens_with_node_type, 'c': self.get_tokens_with_node_type, 'cpp': self.get_tokens_with_node_type, 'c_sharp': self.get_tokens_with_node_type, 'javascript': JavascriptProcessor.get_tokens, 'python': PythonProcessor.get_tokens, 'php': PhpProcessor.get_tokens, 'ruby': self.get_tokens_with_node_type, 'go': self.get_tokens_with_node_type} self.processor = processor_map[language] def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: root_node = self.parse_code(code=code) return_values = self.processor(code=code.encode(), root=root_node) if isinstance(return_values, tuple): (tokens, types) = return_values else: (tokens, types) = (return_values, None) return (re.sub('[ \t\n]+', ' ', ' '.join(tokens)), {'types': types, 'success': False})
class OperandSwap(TransformationBase): '\n Swapping Operand "a>b" becomes "b<a"\n ' def __init__(self, parser_path, language): super(OperandSwap, self).__init__(parser_path=parser_path, language=language) self.language = language self.transformations = processor_function[language] processor_map = {'java': self.get_tokens_with_node_type, 'c': self.get_tokens_with_node_type, 'cpp': self.get_tokens_with_node_type, 'c_sharp': self.get_tokens_with_node_type, 'javascript': JavascriptProcessor.get_tokens, 'python': PythonProcessor.get_tokens, 'php': PhpProcessor.get_tokens, 'ruby': self.get_tokens_with_node_type, 'go': self.get_tokens_with_node_type} self.final_processor = processor_map[self.language] def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: success = False transform_functions = copy.deepcopy(self.transformations) while ((not success) and (len(transform_functions) > 0)): function = np.random.choice(transform_functions) transform_functions.remove(function) (modified_code, success) = function(code, self) if success: code = modified_code root_node = self.parse_code(code=code) return_values = self.final_processor(code=code.encode(), root=root_node) if isinstance(return_values, tuple): (tokens, types) = return_values else: (tokens, types) = (return_values, None) return (re.sub('[ \t\n]+', ' ', ' '.join(tokens)), {'types': types, 'success': success})
def masking(tokens, p): new_tokens = [] for t in tokens: if (np.random.uniform() < p): new_tokens.append('<mask>') else: new_tokens.append(t) return ' '.join(new_tokens)
def deletion(tokens, p): new_tokens = [] for t in tokens: if (np.random.uniform() >= p): new_tokens.append(t) return ' '.join(new_tokens)
def token_infilling(tokens, p): new_tokens = [] max_infilling_len = round((int((p * len(tokens))) / 2.0)) infilling_len = np.random.randint(1, max_infilling_len) start_index = np.random.uniform(high=(len(tokens) - infilling_len)) end_index = (start_index + infilling_len) for (i, t) in enumerate(tokens): if ((i < start_index) or (i > end_index)): new_tokens.append(t) return ' '.join(new_tokens)
class SyntacticNoisingTransformation(TransformationBase): def __init__(self, parser_path: str, language: str, noise_ratio=0.15): self.language = language if (self.language == 'nl'): self.tokenizer = nltk.word_tokenize else: self.tokenizer = NoTransformation(parser_path, language) self.noise_ratio = noise_ratio def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: if (self.language == 'nl'): tokens = self.tokenizer(code) else: (tokenized_code, _) = self.tokenizer.transform_code(code) tokens = tokenized_code.split() p = np.random.uniform() if (p < 0.33): transformed_code = masking(tokens, self.noise_ratio) elif (p < 0.66): transformed_code = deletion(tokens, self.noise_ratio) else: transformed_code = token_infilling(tokens, self.noise_ratio) return (transformed_code, {'success': True})
def get_ancestor_type_chains(node: tree_sitter.Node) -> List[str]: types = [str(node.type)] while (node.parent is not None): node = node.parent types.append(str(node.type)) return types
class TransformationBase(): def __init__(self, parser_path: str, language: str): if (not os.path.exists(parser_path)): raise ValueError(f'Language parser does not exist at {parser_path}. Please run `setup.sh` to properly set the environment!') self.lang_object = Language(parser_path, language) self.parser = Parser() self.parser.set_language(self.lang_object) pass def parse_code(self, code: Union[(str, bytes)]) -> tree_sitter.Node: '\n This function parses a given code and return the root node.\n :param code:\n :return: tree_sitter.Node, the root node of the parsed tree.\n ' if isinstance(code, bytes): tree = self.parser.parse(code) elif isinstance(code, str): tree = self.parser.parse(code.encode()) else: raise ValueError('Code must be character string or bytes string') return tree.root_node def get_tokens(self, code: bytes, root: tree_sitter.Node) -> List[str]: '\n This function is for getting tokens recursively from a tree.\n :param code: the byte string corresponding to the code.\n :param root: the root node of the parsed tree\n :return: List of Tokens.\n ' tokens = [] if (root.type == 'comment'): return tokens if ('string' in str(root.type)): parent = root.parent if (('list' not in str(parent.type)) and (len(parent.children) == 1)): return tokens else: return [code[root.start_byte:root.end_byte].decode()] if (len(root.children) == 0): tokens.append(code[root.start_byte:root.end_byte].decode()) else: for child in root.children: tokens += self.get_tokens(code, child) return tokens def get_token_string(self, code: str, root: tree_sitter.Node) -> str: '\n This is a auxiliary function for just extracting the parsed token string.\n :param code: the byte string corresponding to the code.\n :param root: the root node of the parsed tree\n :return: str, the parsed code a string of tokens.\n ' tokens = self.get_tokens(code.encode(), root) return ' '.join(tokens) def get_tokens_with_node_type(self, code: bytes, root: tree_sitter.Node) -> Tuple[(List[str], List[List[str]])]: '\n This function extracts the tokens and types of the tokens.\n It returns a list of string as tokens, and a list of list of string as types.\n For every token, it extracts the sequence of ast node type starting from the token all the way to the root.\n :param code: the byte string corresponding to the code.\n :param root: the root node of the parsed tree\n :return:\n List[str]: The list of tokens.\n List[List[str]]: The AST node types corresponding to every token.\n ' (tokens, types) = ([], []) if (root.type == 'comment'): return (tokens, types) if ('string' in str(root.type)): return ([code[root.start_byte:root.end_byte].decode()], [['string']]) if (len(root.children) == 0): tokens.append(code[root.start_byte:root.end_byte].decode()) types.append(get_ancestor_type_chains(root)) else: for child in root.children: (_tokens, _types) = self.get_tokens_with_node_type(code, child) tokens += _tokens types += _types return (tokens, types) def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: '\n Transforms a piece of code and returns the transformed version\n :param code: The code to be transformed either as a character string of bytes string.\n :return:\n A tuple, where the first member is the transformed code.\n The second member might be other metadata (e.g. nde types) of the transformed code. It can be None as well.\n ' pass
class SemanticPreservingTransformation(): def __init__(self, parser_path: str, language: str, transform_functions: Dict[(Callable, int)]=None): self.language = language if (transform_functions is not None): self.transform_functions = transform_functions else: self.transform_functions = {BlockSwap: 1, ConfusionRemover: 1, DeadCodeInserter: 1, ForWhileTransformer: 1, OperandSwap: 1, SyntacticNoisingTransformation: 1} self.transformations = [] if (self.language == 'nl'): self.transformations.append(SyntacticNoisingTransformation(parser_path=parser_path, language='nl')) else: for t in self.transform_functions: for _ in range(self.transform_functions[t]): self.transformations.append(t(parser_path=parser_path, language=language)) def transform_code(self, code: str): (transformed_code, transformation_name) = (None, None) indices = list(range(len(self.transformations))) np.random.shuffle(indices) success = False while ((not success) and (len(indices) > 0)): si = np.random.choice(indices) indices.remove(si) t = self.transformations[si] (transformed_code, metadata) = t.transform_code(code) success = metadata['success'] if success: transformation_name = type(t).__name__ if (not success): return (code, None) return (transformed_code, transformation_name)
class VarRenamer(TransformationBase): def __init__(self, parser_path: str, language: str): super(VarRenamer, self).__init__(parser_path=parser_path, language=language) self.language = language self.processor = processor_function[self.language] self.tokenizer_function = tokenizer_function[self.language] self.not_var_ptype = ['function_declarator', 'class_declaration', 'method_declaration', 'function_definition', 'function_declaration', 'call', 'local_function_statement'] def extract_var_names(self, root, code_string): var_names = [] queue = [root] while (len(queue) > 0): current_node = queue[0] queue = queue[1:] if (((current_node.type == 'identifier') or (current_node.type == 'variable_name')) and (str(current_node.parent.type) not in self.not_var_ptype)): var_names.append(self.tokenizer_function(code_string, current_node)[0]) for child in current_node.children: queue.append(child) return var_names def var_renaming(self, code_string): root = self.parse_code(code_string) original_code = self.tokenizer_function(code_string, root) var_names = self.extract_var_names(root, code_string) var_names = list(set(var_names)) num_to_rename = math.ceil((0.2 * len(var_names))) random.shuffle(var_names) var_names = var_names[:num_to_rename] var_map = {} for (idx, v) in enumerate(var_names): var_map[v] = f'VAR_{idx}' modified_code = [] for t in original_code: if (t in var_names): modified_code.append(var_map[t]) else: modified_code.append(t) modified_code_string = ' '.join(modified_code) if (modified_code != original_code): modified_root = self.parse_code(modified_code_string) return (modified_root, modified_code_string, True) else: return (root, code_string, False) def transform_code(self, code: Union[(str, bytes)]) -> Tuple[(str, object)]: (root, code, success) = self.var_renaming(code) code = re.sub('[ \n\t]+', ' ', code) return (code, {'success': success})
def sentence_bleu(references, hypothesis, weights=(0.25, 0.25, 0.25, 0.25), smoothing_function=None, auto_reweigh=False): '\n Calculate BLEU score (Bilingual Evaluation Understudy) from\n Papineni, Kishore, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002.\n "BLEU: a method for automatic evaluation of machine translation."\n In Proceedings of ACL. http://www.aclweb.org/anthology/P02-1040.pdf\n >>> hypothesis1 = [\'It\', \'is\', \'a\', \'guide\', \'to\', \'action\', \'which\',\n ... \'ensures\', \'that\', \'the\', \'military\', \'always\',\n ... \'obeys\', \'the\', \'commands\', \'of\', \'the\', \'party\']\n >>> hypothesis2 = [\'It\', \'is\', \'to\', \'insure\', \'the\', \'troops\',\n ... \'forever\', \'hearing\', \'the\', \'activity\', \'guidebook\',\n ... \'that\', \'party\', \'direct\']\n >>> reference1 = [\'It\', \'is\', \'a\', \'guide\', \'to\', \'action\', \'that\',\n ... \'ensures\', \'that\', \'the\', \'military\', \'will\', \'forever\',\n ... \'heed\', \'Party\', \'commands\']\n >>> reference2 = [\'It\', \'is\', \'the\', \'guiding\', \'principle\', \'which\',\n ... \'guarantees\', \'the\', \'military\', \'forces\', \'always\',\n ... \'being\', \'under\', \'the\', \'command\', \'of\', \'the\',\n ... \'Party\']\n >>> reference3 = [\'It\', \'is\', \'the\', \'practical\', \'guide\', \'for\', \'the\',\n ... \'army\', \'always\', \'to\', \'heed\', \'the\', \'directions\',\n ... \'of\', \'the\', \'party\']\n >>> sentence_bleu([reference1, reference2, reference3], hypothesis1) # doctest: +ELLIPSIS\n 0.5045...\n If there is no ngrams overlap for any order of n-grams, BLEU returns the\n value 0. This is because the precision for the order of n-grams without\n overlap is 0, and the geometric mean in the final BLEU score computation\n multiplies the 0 with the precision of other n-grams. This results in 0\n (independently of the precision of the othe n-gram orders). The following\n example has zero 3-gram and 4-gram overlaps:\n >>> round(sentence_bleu([reference1, reference2, reference3], hypothesis2),4) # doctest: +ELLIPSIS\n 0.0\n To avoid this harsh behaviour when no ngram overlaps are found a smoothing\n function can be used.\n >>> chencherry = SmoothingFunction()\n >>> sentence_bleu([reference1, reference2, reference3], hypothesis2,\n ... smoothing_function=chencherry.method1) # doctest: +ELLIPSIS\n 0.0370...\n The default BLEU calculates a score for up to 4-grams using uniform\n weights (this is called BLEU-4). To evaluate your translations with\n higher/lower order ngrams, use customized weights. E.g. when accounting\n for up to 5-grams with uniform weights (this is called BLEU-5) use:\n >>> weights = (1./5., 1./5., 1./5., 1./5., 1./5.)\n >>> sentence_bleu([reference1, reference2, reference3], hypothesis1, weights) # doctest: +ELLIPSIS\n 0.3920...\n :param references: reference sentences\n :type references: list(list(str))\n :param hypothesis: a hypothesis sentence\n :type hypothesis: list(str)\n :param weights: weights for unigrams, bigrams, trigrams and so on\n :type weights: list(float)\n :param smoothing_function:\n :type smoothing_function: SmoothingFunction\n :param auto_reweigh: Option to re-normalize the weights uniformly.\n :type auto_reweigh: bool\n :return: The sentence-level BLEU score.\n :rtype: float\n ' return corpus_bleu([references], [hypothesis], weights, smoothing_function, auto_reweigh)
def corpus_bleu(list_of_references, hypotheses, weights=(0.25, 0.25, 0.25, 0.25), smoothing_function=None, auto_reweigh=False): "\n Calculate a single corpus-level BLEU score (aka. system-level BLEU) for all\n the hypotheses and their respective references.\n Instead of averaging the sentence level BLEU scores (i.e. marco-average\n precision), the original BLEU metric (Papineni et al. 2002) accounts for\n the micro-average precision (i.e. summing the numerators and denominators\n for each hypothesis-reference(s) pairs before the division).\n >>> hyp1 = ['It', 'is', 'a', 'guide', 'to', 'action', 'which',\n ... 'ensures', 'that', 'the', 'military', 'always',\n ... 'obeys', 'the', 'commands', 'of', 'the', 'party']\n >>> ref1a = ['It', 'is', 'a', 'guide', 'to', 'action', 'that',\n ... 'ensures', 'that', 'the', 'military', 'will', 'forever',\n ... 'heed', 'Party', 'commands']\n >>> ref1b = ['It', 'is', 'the', 'guiding', 'principle', 'which',\n ... 'guarantees', 'the', 'military', 'forces', 'always',\n ... 'being', 'under', 'the', 'command', 'of', 'the', 'Party']\n >>> ref1c = ['It', 'is', 'the', 'practical', 'guide', 'for', 'the',\n ... 'army', 'always', 'to', 'heed', 'the', 'directions',\n ... 'of', 'the', 'party']\n >>> hyp2 = ['he', 'read', 'the', 'book', 'because', 'he', 'was',\n ... 'interested', 'in', 'world', 'history']\n >>> ref2a = ['he', 'was', 'interested', 'in', 'world', 'history',\n ... 'because', 'he', 'read', 'the', 'book']\n >>> list_of_references = [[ref1a, ref1b, ref1c], [ref2a]]\n >>> hypotheses = [hyp1, hyp2]\n >>> corpus_bleu(list_of_references, hypotheses) # doctest: +ELLIPSIS\n 0.5920...\n The example below show that corpus_bleu() is different from averaging\n sentence_bleu() for hypotheses\n >>> score1 = sentence_bleu([ref1a, ref1b, ref1c], hyp1)\n >>> score2 = sentence_bleu([ref2a], hyp2)\n >>> (score1 + score2) / 2 # doctest: +ELLIPSIS\n 0.6223...\n :param list_of_references: a corpus of lists of reference sentences, w.r.t. hypotheses\n :type list_of_references: list(list(list(str)))\n :param hypotheses: a list of hypothesis sentences\n :type hypotheses: list(list(str))\n :param weights: weights for unigrams, bigrams, trigrams and so on\n :type weights: list(float)\n :param smoothing_function:\n :type smoothing_function: SmoothingFunction\n :param auto_reweigh: Option to re-normalize the weights uniformly.\n :type auto_reweigh: bool\n :return: The corpus-level BLEU score.\n :rtype: float\n " p_numerators = Counter() p_denominators = Counter() (hyp_lengths, ref_lengths) = (0, 0) assert (len(list_of_references) == len(hypotheses)), 'The number of hypotheses and their reference(s) should be the same ' for (references, hypothesis) in zip(list_of_references, hypotheses): for (i, _) in enumerate(weights, start=1): p_i = modified_precision(references, hypothesis, i) p_numerators[i] += p_i.numerator p_denominators[i] += p_i.denominator hyp_len = len(hypothesis) hyp_lengths += hyp_len ref_lengths += closest_ref_length(references, hyp_len) bp = brevity_penalty(ref_lengths, hyp_lengths) if auto_reweigh: if ((hyp_lengths < 4) and (weights == (0.25, 0.25, 0.25, 0.25))): weights = (((1 / hyp_lengths),) * hyp_lengths) p_n = [Fraction(p_numerators[i], p_denominators[i], _normalize=False) for (i, _) in enumerate(weights, start=1)] if (p_numerators[1] == 0): return 0 if (not smoothing_function): smoothing_function = SmoothingFunction().method1 p_n = smoothing_function(p_n, references=references, hypothesis=hypothesis, hyp_len=hyp_lengths) s = ((w_i * math.log(p_i)) for (w_i, p_i) in zip(weights, p_n)) s = (bp * math.exp(math.fsum(s))) return s
def modified_precision(references, hypothesis, n): '\n Calculate modified ngram precision.\n The normal precision method may lead to some wrong translations with\n high-precision, e.g., the translation, in which a word of reference\n repeats several times, has very high precision.\n This function only returns the Fraction object that contains the numerator\n and denominator necessary to calculate the corpus-level precision.\n To calculate the modified precision for a single pair of hypothesis and\n references, cast the Fraction object into a float.\n The famous "the the the ... " example shows that you can get BLEU precision\n by duplicating high frequency words.\n >>> reference1 = \'the cat is on the mat\'.split()\n >>> reference2 = \'there is a cat on the mat\'.split()\n >>> hypothesis1 = \'the the the the the the the\'.split()\n >>> references = [reference1, reference2]\n >>> float(modified_precision(references, hypothesis1, n=1)) # doctest: +ELLIPSIS\n 0.2857...\n In the modified n-gram precision, a reference word will be considered\n exhausted after a matching hypothesis word is identified, e.g.\n >>> reference1 = [\'It\', \'is\', \'a\', \'guide\', \'to\', \'action\', \'that\',\n ... \'ensures\', \'that\', \'the\', \'military\', \'will\',\n ... \'forever\', \'heed\', \'Party\', \'commands\']\n >>> reference2 = [\'It\', \'is\', \'the\', \'guiding\', \'principle\', \'which\',\n ... \'guarantees\', \'the\', \'military\', \'forces\', \'always\',\n ... \'being\', \'under\', \'the\', \'command\', \'of\', \'the\',\n ... \'Party\']\n >>> reference3 = [\'It\', \'is\', \'the\', \'practical\', \'guide\', \'for\', \'the\',\n ... \'army\', \'always\', \'to\', \'heed\', \'the\', \'directions\',\n ... \'of\', \'the\', \'party\']\n >>> hypothesis = \'of the\'.split()\n >>> references = [reference1, reference2, reference3]\n >>> float(modified_precision(references, hypothesis, n=1))\n 1.0\n >>> float(modified_precision(references, hypothesis, n=2))\n 1.0\n An example of a normal machine translation hypothesis:\n >>> hypothesis1 = [\'It\', \'is\', \'a\', \'guide\', \'to\', \'action\', \'which\',\n ... \'ensures\', \'that\', \'the\', \'military\', \'always\',\n ... \'obeys\', \'the\', \'commands\', \'of\', \'the\', \'party\']\n >>> hypothesis2 = [\'It\', \'is\', \'to\', \'insure\', \'the\', \'troops\',\n ... \'forever\', \'hearing\', \'the\', \'activity\', \'guidebook\',\n ... \'that\', \'party\', \'direct\']\n >>> reference1 = [\'It\', \'is\', \'a\', \'guide\', \'to\', \'action\', \'that\',\n ... \'ensures\', \'that\', \'the\', \'military\', \'will\',\n ... \'forever\', \'heed\', \'Party\', \'commands\']\n >>> reference2 = [\'It\', \'is\', \'the\', \'guiding\', \'principle\', \'which\',\n ... \'guarantees\', \'the\', \'military\', \'forces\', \'always\',\n ... \'being\', \'under\', \'the\', \'command\', \'of\', \'the\',\n ... \'Party\']\n >>> reference3 = [\'It\', \'is\', \'the\', \'practical\', \'guide\', \'for\', \'the\',\n ... \'army\', \'always\', \'to\', \'heed\', \'the\', \'directions\',\n ... \'of\', \'the\', \'party\']\n >>> references = [reference1, reference2, reference3]\n >>> float(modified_precision(references, hypothesis1, n=1)) # doctest: +ELLIPSIS\n 0.9444...\n >>> float(modified_precision(references, hypothesis2, n=1)) # doctest: +ELLIPSIS\n 0.5714...\n >>> float(modified_precision(references, hypothesis1, n=2)) # doctest: +ELLIPSIS\n 0.5882352941176471\n >>> float(modified_precision(references, hypothesis2, n=2)) # doctest: +ELLIPSIS\n 0.07692...\n :param references: A list of reference translations.\n :type references: list(list(str))\n :param hypothesis: A hypothesis translation.\n :type hypothesis: list(str)\n :param n: The ngram order.\n :type n: int\n :return: BLEU\'s modified precision for the nth order ngram.\n :rtype: Fraction\n ' counts = (Counter(ngrams(hypothesis, n)) if (len(hypothesis) >= n) else Counter()) max_counts = {} for reference in references: reference_counts = (Counter(ngrams(reference, n)) if (len(reference) >= n) else Counter()) for ngram in counts: max_counts[ngram] = max(max_counts.get(ngram, 0), reference_counts[ngram]) clipped_counts = {ngram: min(count, max_counts[ngram]) for (ngram, count) in counts.items()} numerator = sum(clipped_counts.values()) denominator = max(1, sum(counts.values())) return Fraction(numerator, denominator, _normalize=False)
def closest_ref_length(references, hyp_len): "\n This function finds the reference that is the closest length to the\n hypothesis. The closest reference length is referred to as *r* variable\n from the brevity penalty formula in Papineni et. al. (2002)\n :param references: A list of reference translations.\n :type references: list(list(str))\n :param hyp_len: The length of the hypothesis.\n :type hyp_len: int\n :return: The length of the reference that's closest to the hypothesis.\n :rtype: int\n " ref_lens = (len(reference) for reference in references) closest_ref_len = min(ref_lens, key=(lambda ref_len: (abs((ref_len - hyp_len)), ref_len))) return closest_ref_len
def brevity_penalty(closest_ref_len, hyp_len): "\n Calculate brevity penalty.\n As the modified n-gram precision still has the problem from the short\n length sentence, brevity penalty is used to modify the overall BLEU\n score according to length.\n An example from the paper. There are three references with length 12, 15\n and 17. And a concise hypothesis of the length 12. The brevity penalty is 1.\n >>> reference1 = list('aaaaaaaaaaaa') # i.e. ['a'] * 12\n >>> reference2 = list('aaaaaaaaaaaaaaa') # i.e. ['a'] * 15\n >>> reference3 = list('aaaaaaaaaaaaaaaaa') # i.e. ['a'] * 17\n >>> hypothesis = list('aaaaaaaaaaaa') # i.e. ['a'] * 12\n >>> references = [reference1, reference2, reference3]\n >>> hyp_len = len(hypothesis)\n >>> closest_ref_len = closest_ref_length(references, hyp_len)\n >>> brevity_penalty(closest_ref_len, hyp_len)\n 1.0\n In case a hypothesis translation is shorter than the references, penalty is\n applied.\n >>> references = [['a'] * 28, ['a'] * 28]\n >>> hypothesis = ['a'] * 12\n >>> hyp_len = len(hypothesis)\n >>> closest_ref_len = closest_ref_length(references, hyp_len)\n >>> brevity_penalty(closest_ref_len, hyp_len)\n 0.2635971381157267\n The length of the closest reference is used to compute the penalty. If the\n length of a hypothesis is 12, and the reference lengths are 13 and 2, the\n penalty is applied because the hypothesis length (12) is less then the\n closest reference length (13).\n >>> references = [['a'] * 13, ['a'] * 2]\n >>> hypothesis = ['a'] * 12\n >>> hyp_len = len(hypothesis)\n >>> closest_ref_len = closest_ref_length(references, hyp_len)\n >>> brevity_penalty(closest_ref_len, hyp_len) # doctest: +ELLIPSIS\n 0.9200...\n The brevity penalty doesn't depend on reference order. More importantly,\n when two reference sentences are at the same distance, the shortest\n reference sentence length is used.\n >>> references = [['a'] * 13, ['a'] * 11]\n >>> hypothesis = ['a'] * 12\n >>> hyp_len = len(hypothesis)\n >>> closest_ref_len = closest_ref_length(references, hyp_len)\n >>> bp1 = brevity_penalty(closest_ref_len, hyp_len)\n >>> hyp_len = len(hypothesis)\n >>> closest_ref_len = closest_ref_length(reversed(references), hyp_len)\n >>> bp2 = brevity_penalty(closest_ref_len, hyp_len)\n >>> bp1 == bp2 == 1\n True\n A test example from mteval-v13a.pl (starting from the line 705):\n >>> references = [['a'] * 11, ['a'] * 8]\n >>> hypothesis = ['a'] * 7\n >>> hyp_len = len(hypothesis)\n >>> closest_ref_len = closest_ref_length(references, hyp_len)\n >>> brevity_penalty(closest_ref_len, hyp_len) # doctest: +ELLIPSIS\n 0.8668...\n >>> references = [['a'] * 11, ['a'] * 8, ['a'] * 6, ['a'] * 7]\n >>> hypothesis = ['a'] * 7\n >>> hyp_len = len(hypothesis)\n >>> closest_ref_len = closest_ref_length(references, hyp_len)\n >>> brevity_penalty(closest_ref_len, hyp_len)\n 1.0\n :param hyp_len: The length of the hypothesis for a single sentence OR the\n sum of all the hypotheses' lengths for a corpus\n :type hyp_len: int\n :param closest_ref_len: The length of the closest reference for a single\n hypothesis OR the sum of all the closest references for every hypotheses.\n :type closest_ref_len: int\n :return: BLEU's brevity penalty.\n :rtype: float\n " if (hyp_len > closest_ref_len): return 1 elif (hyp_len == 0): return 0 else: return math.exp((1 - (closest_ref_len / hyp_len)))
class SmoothingFunction(): '\n This is an implementation of the smoothing techniques\n for segment-level BLEU scores that was presented in\n Boxing Chen and Collin Cherry (2014) A Systematic Comparison of\n Smoothing Techniques for Sentence-Level BLEU. In WMT14.\n http://acl2014.org/acl2014/W14-33/pdf/W14-3346.pdf\n ' def __init__(self, epsilon=0.1, alpha=5, k=5): "\n This will initialize the parameters required for the various smoothing\n techniques, the default values are set to the numbers used in the\n experiments from Chen and Cherry (2014).\n >>> hypothesis1 = ['It', 'is', 'a', 'guide', 'to', 'action', 'which', 'ensures',\n ... 'that', 'the', 'military', 'always', 'obeys', 'the',\n ... 'commands', 'of', 'the', 'party']\n >>> reference1 = ['It', 'is', 'a', 'guide', 'to', 'action', 'that', 'ensures',\n ... 'that', 'the', 'military', 'will', 'forever', 'heed',\n ... 'Party', 'commands']\n >>> chencherry = SmoothingFunction()\n >>> print(sentence_bleu([reference1], hypothesis1)) # doctest: +ELLIPSIS\n 0.4118...\n >>> print(sentence_bleu([reference1], hypothesis1, smoothing_function=chencherry.method0)) # doctest: +ELLIPSIS\n 0.4118...\n >>> print(sentence_bleu([reference1], hypothesis1, smoothing_function=chencherry.method1)) # doctest: +ELLIPSIS\n 0.4118...\n >>> print(sentence_bleu([reference1], hypothesis1, smoothing_function=chencherry.method2)) # doctest: +ELLIPSIS\n 0.4489...\n >>> print(sentence_bleu([reference1], hypothesis1, smoothing_function=chencherry.method3)) # doctest: +ELLIPSIS\n 0.4118...\n >>> print(sentence_bleu([reference1], hypothesis1, smoothing_function=chencherry.method4)) # doctest: +ELLIPSIS\n 0.4118...\n >>> print(sentence_bleu([reference1], hypothesis1, smoothing_function=chencherry.method5)) # doctest: +ELLIPSIS\n 0.4905...\n >>> print(sentence_bleu([reference1], hypothesis1, smoothing_function=chencherry.method6)) # doctest: +ELLIPSIS\n 0.4135...\n >>> print(sentence_bleu([reference1], hypothesis1, smoothing_function=chencherry.method7)) # doctest: +ELLIPSIS\n 0.4905...\n :param epsilon: the epsilon value use in method 1\n :type epsilon: float\n :param alpha: the alpha value use in method 6\n :type alpha: int\n :param k: the k value use in method 4\n :type k: int\n " self.epsilon = epsilon self.alpha = alpha self.k = k def method0(self, p_n, *args, **kwargs): '\n No smoothing.\n ' p_n_new = [] for (i, p_i) in enumerate(p_n): if (p_i.numerator != 0): p_n_new.append(p_i) else: _msg = str('\nThe hypothesis contains 0 counts of {}-gram overlaps.\nTherefore the BLEU score evaluates to 0, independently of\nhow many N-gram overlaps of lower order it contains.\nConsider using lower n-gram order or use SmoothingFunction()').format((i + 1)) warnings.warn(_msg) p_n_new.append(sys.float_info.min) return p_n_new def method1(self, p_n, *args, **kwargs): '\n Smoothing method 1: Add *epsilon* counts to precision with 0 counts.\n ' return [(((p_i.numerator + self.epsilon) / p_i.denominator) if (p_i.numerator == 0) else p_i) for p_i in p_n] def method2(self, p_n, *args, **kwargs): '\n Smoothing method 2: Add 1 to both numerator and denominator from\n Chin-Yew Lin and Franz Josef Och (2004) Automatic evaluation of\n machine translation quality using longest common subsequence and\n skip-bigram statistics. In ACL04.\n ' return [Fraction((p_i.numerator + 1), (p_i.denominator + 1), _normalize=False) for p_i in p_n] def method3(self, p_n, *args, **kwargs): "\n Smoothing method 3: NIST geometric sequence smoothing\n The smoothing is computed by taking 1 / ( 2^k ), instead of 0, for each\n precision score whose matching n-gram count is null.\n k is 1 for the first 'n' value for which the n-gram match count is null/\n For example, if the text contains:\n - one 2-gram match\n - and (consequently) two 1-gram matches\n the n-gram count for each individual precision score would be:\n - n=1 => prec_count = 2 (two unigrams)\n - n=2 => prec_count = 1 (one bigram)\n - n=3 => prec_count = 1/2 (no trigram, taking 'smoothed' value of 1 / ( 2^k ), with k=1)\n - n=4 => prec_count = 1/4 (no fourgram, taking 'smoothed' value of 1 / ( 2^k ), with k=2)\n " incvnt = 1 for (i, p_i) in enumerate(p_n): if (p_i.numerator == 0): p_n[i] = (1 / ((2 ** incvnt) * p_i.denominator)) incvnt += 1 return p_n def method4(self, p_n, references, hypothesis, hyp_len=None, *args, **kwargs): '\n Smoothing method 4:\n Shorter translations may have inflated precision values due to having\n smaller denominators; therefore, we give them proportionally\n smaller smoothed counts. Instead of scaling to 1/(2^k), Chen and Cherry\n suggests dividing by 1/ln(len(T)), where T is the length of the translation.\n ' hyp_len = (hyp_len if hyp_len else len(hypothesis)) for (i, p_i) in enumerate(p_n): if ((p_i.numerator == 0) and (hyp_len != 0)): incvnt = (i + ((1 * self.k) / math.log(hyp_len))) p_n[i] = (incvnt / p_i.denominator) return p_n def method5(self, p_n, references, hypothesis, hyp_len=None, *args, **kwargs): '\n Smoothing method 5:\n The matched counts for similar values of n should be similar. To a\n calculate the n-gram matched count, it averages the n−1, n and n+1 gram\n matched counts.\n ' hyp_len = (hyp_len if hyp_len else len(hypothesis)) m = {} p_n_plus1 = (p_n + [modified_precision(references, hypothesis, 5)]) m[(- 1)] = (p_n[0] + 1) for (i, p_i) in enumerate(p_n): p_n[i] = (((m[(i - 1)] + p_i) + p_n_plus1[(i + 1)]) / 3) m[i] = p_n[i] return p_n def method6(self, p_n, references, hypothesis, hyp_len=None, *args, **kwargs): '\n Smoothing method 6:\n Interpolates the maximum likelihood estimate of the precision *p_n* with\n a prior estimate *pi0*. The prior is estimated by assuming that the ratio\n between pn and pn−1 will be the same as that between pn−1 and pn−2; from\n Gao and He (2013) Training MRF-Based Phrase Translation Models using\n Gradient Ascent. In NAACL.\n ' hyp_len = (hyp_len if hyp_len else len(hypothesis)) assert p_n[2], 'This smoothing method requires non-zero precision for bigrams.' for (i, p_i) in enumerate(p_n): if (i in [0, 1]): continue else: pi0 = (0 if (p_n[(i - 2)] == 0) else ((p_n[(i - 1)] ** 2) / p_n[(i - 2)])) m = p_i.numerator l = sum((1 for _ in ngrams(hypothesis, (i + 1)))) p_n[i] = ((m + (self.alpha * pi0)) / (l + self.alpha)) return p_n def method7(self, p_n, references, hypothesis, hyp_len=None, *args, **kwargs): '\n Smoothing method 7:\n Interpolates methods 4 and 5.\n ' hyp_len = (hyp_len if hyp_len else len(hypothesis)) p_n = self.method4(p_n, references, hypothesis, hyp_len) p_n = self.method5(p_n, references, hypothesis, hyp_len) return p_n
def evaluate_per_example(reference, hypothesis, lang, params='0.25,0.25,0.25,0.25'): (alpha, beta, gamma, theta) = [float(x) for x in params.split(',')] hypothesis = [hypothesis] pre_references = [[reference]] for i in range(len(pre_references)): assert (len(hypothesis) == len(pre_references[i])) references = [] for i in range(len(hypothesis)): ref_for_instance = [] for j in range(len(pre_references)): ref_for_instance.append(pre_references[j][i]) references.append(ref_for_instance) assert (len(references) == (len(pre_references) * len(hypothesis))) tokenized_hyps = [x.split() for x in hypothesis] tokenized_refs = [[x.split() for x in reference] for reference in references] ngram_match_score = bleu.corpus_bleu(tokenized_refs, tokenized_hyps) root_dir = os.path.dirname(__file__) keywords = [x.strip() for x in open((((root_dir + '/keywords/') + lang) + '.txt'), 'r', encoding='utf-8').readlines()] def make_weights(reference_tokens, key_word_list): return {token: (1 if (token in key_word_list) else 0.2) for token in reference_tokens} tokenized_refs_with_weights = [[[reference_tokens, make_weights(reference_tokens, keywords)] for reference_tokens in reference] for reference in tokenized_refs] weighted_ngram_match_score = weighted_ngram_match.corpus_bleu(tokenized_refs_with_weights, tokenized_hyps) syntax_match_score = syntax_match.corpus_syntax_match(references, hypothesis, lang) dataflow_match_score = dataflow_match.corpus_dataflow_match(references, hypothesis, lang) print('ngram match: {0}, weighted ngram match: {1}, syntax_match: {2}, dataflow_match: {3}'.format(ngram_match_score, weighted_ngram_match_score, syntax_match_score, dataflow_match_score)) codebleu = ((((alpha * ngram_match_score) + (beta * weighted_ngram_match_score)) + (gamma * syntax_match_score)) + (theta * dataflow_match_score)) return {'em': (1.0 if (reference.strip() == hypothesis[0].strip()) else 0.0), 'bleu': ngram_match_score, 'wbleu': weighted_ngram_match_score, 'syntax': syntax_match_score, 'dataflow': dataflow_match_score, 'codebleu': codebleu}
def get_codebleu(refs, hyp, lang, params='0.25,0.25,0.25,0.25'): if (not isinstance(refs, list)): refs = [refs] (alpha, beta, gamma, theta) = [float(x) for x in params.split(',')] pre_references = [[x.strip() for x in open(file, 'r', encoding='utf-8').readlines()] for file in refs] hypothesis = [x.strip() for x in open(hyp, 'r', encoding='utf-8').readlines()] for i in range(len(pre_references)): assert (len(hypothesis) == len(pre_references[i])) references = [] for i in range(len(hypothesis)): ref_for_instance = [] for j in range(len(pre_references)): ref_for_instance.append(pre_references[j][i]) references.append(ref_for_instance) assert (len(references) == (len(pre_references) * len(hypothesis))) tokenized_hyps = [x.split() for x in hypothesis] tokenized_refs = [[x.split() for x in reference] for reference in references] ngram_match_score = bleu.corpus_bleu(tokenized_refs, tokenized_hyps) root_dir = os.path.dirname(__file__) keywords = [x.strip() for x in open((((root_dir + '/keywords/') + lang) + '.txt'), 'r', encoding='utf-8').readlines()] def make_weights(reference_tokens, key_word_list): return {token: (1 if (token in key_word_list) else 0.2) for token in reference_tokens} tokenized_refs_with_weights = [[[reference_tokens, make_weights(reference_tokens, keywords)] for reference_tokens in reference] for reference in tokenized_refs] weighted_ngram_match_score = weighted_ngram_match.corpus_bleu(tokenized_refs_with_weights, tokenized_hyps) syntax_match_score = syntax_match.corpus_syntax_match(references, hypothesis, lang) dataflow_match_score = dataflow_match.corpus_dataflow_match(references, hypothesis, lang) print('ngram match: {0}, weighted ngram match: {1}, syntax_match: {2}, dataflow_match: {3}'.format(ngram_match_score, weighted_ngram_match_score, syntax_match_score, dataflow_match_score)) codebleu = ((((alpha * ngram_match_score) + (beta * weighted_ngram_match_score)) + (gamma * syntax_match_score)) + (theta * dataflow_match_score)) return codebleu
def my_dataflow_match(references, candidates, lang): LANGUAGE = Language((root_dir + '/parser/languages.so'), lang) parser = Parser() parser.set_language(LANGUAGE) parser = [parser, dfg_function[lang]] match_count = 0 total_count = 0 candidate_scores = [] for i in range(len(candidates)): scores = [] references_sample = references[i] candidate = candidates[i] for reference in references_sample: try: candidate = remove_comments_and_docstrings(candidate, 'java') except: pass try: reference = remove_comments_and_docstrings(reference, 'java') except: pass cand_dfg = get_data_flow(candidate, parser) ref_dfg = get_data_flow(reference, parser) normalized_cand_dfg = normalize_dataflow(cand_dfg) normalized_ref_dfg = normalize_dataflow(ref_dfg) if (len(normalized_ref_dfg) > 0): total_count += len(normalized_ref_dfg) current_match_count = 0 for dataflow in normalized_ref_dfg: if (dataflow in normalized_cand_dfg): match_count += 1 normalized_cand_dfg.remove(dataflow) current_match_count += 1 scores.append((float(current_match_count) / len(normalized_ref_dfg))) else: scores.append(0.0) candidate_scores.append((max(scores) if (len(scores) > 0) else 0.0)) return (np.mean(candidate_scores) if (len(candidates) > 0) else 0.0)
def calc_dataflow_match(references, candidate, lang): return corpus_dataflow_match([references], [candidate], lang)