processing_moss_audio.py

import re
import types
from dataclasses import dataclass
from typing import List, Optional, Sequence, Union

import numpy as np
import torch
from transformers import BatchFeature
from transformers.processing_utils import ProcessorMixin
from transformers.models.whisper.feature_extraction_whisper import (
    WhisperFeatureExtractor,
)


@dataclass
class MelConfig:
    mel_sr: int = 16000
    mel_dim: int = 128
    mel_n_fft: int = 400
    mel_hop_length: int = 160
    mel_dtype: torch.dtype = torch.bfloat16
    use_whisper_feature_extractor: bool = True


def _normalize_mel_config(mel_config) -> dict[str, object]:
    default_config = MelConfig()
    if mel_config is None:
        source = {}
    elif isinstance(mel_config, MelConfig):
        source = {
            key: getattr(mel_config, key) for key in MelConfig.__dataclass_fields__.keys()
        }
    else:
        source = dict(mel_config)

    normalized = {}
    for key in MelConfig.__dataclass_fields__.keys():
        value = source.get(key, getattr(default_config, key))
        if key == "mel_dtype":
            if isinstance(value, torch.dtype):
                value = str(value).removeprefix("torch.")
            elif isinstance(value, str) and value.startswith("torch."):
                value = value.removeprefix("torch.")
        normalized[key] = value
    return normalized


def _build_mel_config(mel_config_dict: dict[str, object]) -> MelConfig:
    default_config = MelConfig()

    def _int_value(key: str, default: int) -> int:
        value = mel_config_dict.get(key, default)
        if isinstance(value, bool):
            return int(value)
        if isinstance(value, (int, str)):
            return int(value)
        return default

    def _bool_value(key: str, default: bool) -> bool:
        value = mel_config_dict.get(key, default)
        if isinstance(value, bool):
            return value
        if isinstance(value, str):
            return value.lower() in {"1", "true", "yes", "on"}
        if isinstance(value, int):
            return bool(value)
        return default

    mel_dtype_value = mel_config_dict.get("mel_dtype", default_config.mel_dtype)
    if isinstance(mel_dtype_value, str):
        mel_dtype = getattr(torch, mel_dtype_value.removeprefix("torch."))
    elif isinstance(mel_dtype_value, torch.dtype):
        mel_dtype = mel_dtype_value
    else:
        mel_dtype = default_config.mel_dtype

    return MelConfig(
        mel_sr=_int_value("mel_sr", default_config.mel_sr),
        mel_dim=_int_value("mel_dim", default_config.mel_dim),
        mel_n_fft=_int_value("mel_n_fft", default_config.mel_n_fft),
        mel_hop_length=_int_value("mel_hop_length", default_config.mel_hop_length),
        mel_dtype=mel_dtype,
        use_whisper_feature_extractor=_bool_value(
            "use_whisper_feature_extractor",
            default_config.use_whisper_feature_extractor,
        ),
    )


class MossAudioProcessor(ProcessorMixin):
    attributes = ["tokenizer"]
    tokenizer_class = ("Qwen2Tokenizer", "Qwen2TokenizerFast")

    _AUDIO_SPAN_RE = re.compile(r"<\|audio_bos\|>(?:<\|AUDIO\|>)+<\|audio_eos\|>")

    def __init__(
        self,
        tokenizer=None,
        mel_config=None,
        enable_time_marker: bool = False,
        audio_token_id: int = 151654,
        audio_start_id: int = 151669,
        audio_end_id: int = 151670,
        chat_template=None,
    ):
        super().__init__(tokenizer, chat_template=chat_template)
        if tokenizer is None:
            raise ValueError("MossAudioProcessor requires a tokenizer.")

        self._base_tokenizer = tokenizer
        self.mel_config = _normalize_mel_config(mel_config)
        self.config = _build_mel_config(self.mel_config)
        self.enable_time_marker = bool(enable_time_marker)
        self.audio_token_id = int(audio_token_id)
        self.audio_start_id = int(audio_start_id)
        self.audio_end_id = int(audio_end_id)
        self._whisper_feature_extractor = None

        alias_map = {
            "<|AUDIO|>": self.audio_token_id,
            "<|audio_bos|>": self.audio_start_id,
            "<|audio_eos|>": self.audio_end_id,
        }
        orig_convert_tokens_to_ids = tokenizer.convert_tokens_to_ids

        def _patched_convert_tokens_to_ids(tokenizer_self, tokens):
            if isinstance(tokens, (list, tuple)):
                converted = [
                    _patched_convert_tokens_to_ids(tokenizer_self, token)
                    for token in tokens
                ]
                return converted if isinstance(tokens, list) else tuple(converted)
            if isinstance(tokens, str) and tokens in alias_map:
                return alias_map[tokens]
            return orig_convert_tokens_to_ids(tokens)

        tokenizer.convert_tokens_to_ids = types.MethodType(
            _patched_convert_tokens_to_ids, tokenizer
        )

        self._digit_token_ids = {
            "0": 15,
            "1": 16,
            "2": 17,
            "3": 18,
            "4": 19,
            "5": 20,
            "6": 21,
            "7": 22,
            "8": 23,
            "9": 24,
        }
        self.audio_tokens_per_second = 12.5
        self.time_marker_every_seconds = 2
        self.time_marker_every_audio_tokens = int(
            self.audio_tokens_per_second * self.time_marker_every_seconds
        )

    @property
    def model_input_names(self):
        return [
            "input_ids",
            "attention_mask",
            "audio_data",
            "audio_data_seqlens",
        ]

    @staticmethod
    def _conv3_downsample_len(raw_mel_len: int) -> int:
        def conv_out_len(length: int) -> int:
            return (length - 1) // 2 + 1

        length1 = conv_out_len(int(raw_mel_len))
        length2 = conv_out_len(length1)
        length3 = conv_out_len(length2)
        return int(length3)

    def _get_whisper_feature_extractor(self):
        if self._whisper_feature_extractor is not None:
            return self._whisper_feature_extractor

        self._whisper_feature_extractor = WhisperFeatureExtractor(
            feature_size=int(self.config.mel_dim),
            sampling_rate=int(self.config.mel_sr),
            hop_length=int(self.config.mel_hop_length),
            n_fft=int(self.config.mel_n_fft),
        )
        return self._whisper_feature_extractor

    def _extract_mel(self, audio: Union[np.ndarray, torch.Tensor]) -> torch.Tensor:
        if isinstance(audio, np.ndarray):
            wav = torch.from_numpy(audio)
        else:
            wav = audio
        wav = wav.to(dtype=torch.float32)
        if wav.dim() == 1:
            wav = wav.unsqueeze(0)

        if bool(getattr(self.config, "use_whisper_feature_extractor", False)):
            fe = self._get_whisper_feature_extractor()
            wav_np = wav.detach().to("cpu", torch.float32).contiguous().numpy()
            if wav_np.ndim == 2:
                wav_np = wav_np[0]
            feats = fe._np_extract_fbank_features(wav_np[None, ...], device="cpu")
            mel = torch.from_numpy(feats[0])
        else:
            raise ValueError("MossAudioProcessor requires whisper feature extraction.")

        return mel.to(dtype=self.config.mel_dtype)

    def _get_time_marker_token_ids(self, second: int) -> List[int]:
        return [self._digit_token_ids[digit] for digit in str(second)]

    def _build_audio_tokens_with_time_markers(self, audio_seq_len: int) -> List[int]:
        total_duration_seconds = audio_seq_len / self.audio_tokens_per_second
        num_full_seconds = int(total_duration_seconds)

        token_ids: List[int] = []
        audio_tokens_consumed = 0
        for second in range(
            self.time_marker_every_seconds,
            num_full_seconds + 1,
            self.time_marker_every_seconds,
        ):
            marker_pos = (
                second // self.time_marker_every_seconds
            ) * self.time_marker_every_audio_tokens
            audio_segment_len = marker_pos - audio_tokens_consumed
            if audio_segment_len > 0:
                token_ids.extend([self.audio_token_id] * audio_segment_len)
                audio_tokens_consumed += audio_segment_len
            token_ids.extend(self._get_time_marker_token_ids(second))

        remaining = audio_seq_len - audio_tokens_consumed
        if remaining > 0:
            token_ids.extend([self.audio_token_id] * remaining)
        return token_ids

    def _build_audio_placeholder_ids(self, num_audio_tokens: int) -> List[int]:
        if self.enable_time_marker:
            return self._build_audio_tokens_with_time_markers(num_audio_tokens)
        return [self.audio_token_id] * num_audio_tokens

    def _build_default_prompt(self, text: str, has_audio: bool) -> str:
        if has_audio:
            return (
                "<|im_start|>system\n"
                "You are a helpful assistant.<|im_end|>\n"
                "<|im_start|>user\n"
                "<|audio_bos|><|AUDIO|><|audio_eos|>\n"
                f"{text}<|im_end|>\n"
                "<|im_start|>assistant\n"
            )
        return (
            "<|im_start|>system\n"
            "You are a helpful assistant.<|im_end|>\n"
            "<|im_start|>user\n"
            f"{text}<|im_end|>\n"
            "<|im_start|>assistant\n"
        )

    def _build_input_from_prompt(self, prompt: str, token_lens: List[int]) -> List[int]:
        spans = list(self._AUDIO_SPAN_RE.finditer(prompt))
        if len(spans) != len(token_lens):
            raise ValueError(
                f"Audio placeholder count mismatch: found {len(spans)} spans in text, "
                f"but got {len(token_lens)} audio inputs."
            )

        input_ids: List[int] = []
        cursor = 0
        for index, match in enumerate(spans):
            prefix = prompt[cursor : match.start()]
            if prefix:
                input_ids.extend(
                    self._base_tokenizer.encode(prefix, add_special_tokens=False)
                )

            input_ids.append(self.audio_start_id)
            input_ids.extend(self._build_audio_placeholder_ids(int(token_lens[index])))
            input_ids.append(self.audio_end_id)
            cursor = match.end()

        suffix = prompt[cursor:]
        if suffix:
            input_ids.extend(
                self._base_tokenizer.encode(suffix, add_special_tokens=False)
            )
        return input_ids

    def __call__(
        self,
        *args,
        text: Union[str, Sequence[str], None] = None,
        audios: Optional[Sequence[Union[np.ndarray, torch.Tensor]]] = None,
        audio: Optional[Sequence[Union[np.ndarray, torch.Tensor]]] = None,
        return_tensors: str = "pt",
        **kwargs,
    ) -> BatchFeature:
        _ = args, kwargs

        if isinstance(text, str):
            prompt_text: Optional[str] = text
        elif isinstance(text, (list, tuple)):
            if len(text) != 1:
                raise ValueError(f"Expected text batch size 1, got {len(text)}")
            prompt_text = text[0]
            if not isinstance(prompt_text, str):
                raise TypeError("Expected text batch size 1 with string content.")
        elif text is None:
            prompt_text = None
        else:
            raise TypeError("MossAudioProcessor text must be a string or a batch of one string.")

        audio_list = audios if audios is not None else audio
        audio_list = [] if audio_list is None else list(audio_list)

        mels: List[torch.Tensor] = []
        raw_lengths: List[int] = []
        token_lens: List[int] = []
        for one_audio in audio_list:
            mel = self._extract_mel(one_audio)
            raw_len = int(mel.shape[-1])
            mels.append(mel)
            raw_lengths.append(raw_len)
            token_lens.append(self._conv3_downsample_len(raw_len))

        if mels:
            max_length = max(raw_lengths)
            audio_batch = torch.zeros(
                (len(mels), self.config.mel_dim, max_length),
                dtype=self.config.mel_dtype,
            )
            for index, mel in enumerate(mels):
                audio_batch[index, :, : mel.shape[-1]] = mel
            seqlens_tensor = torch.tensor(raw_lengths, dtype=torch.long)
        else:
            audio_batch = None
            seqlens_tensor = None

        if prompt_text is None:
            raise ValueError(
                "MossAudioProcessor requires text input. Apply a chat template before calling the processor if needed."
            )

        if self._AUDIO_SPAN_RE.search(prompt_text) is None and audio_list:
            prompt_text = self._build_default_prompt(prompt_text, has_audio=True)
        elif self._AUDIO_SPAN_RE.search(prompt_text) is None and not audio_list:
            prompt_text = self._build_default_prompt(prompt_text, has_audio=False)
        input_ids_list = self._build_input_from_prompt(prompt_text, token_lens)

        input_ids_tensor = torch.tensor([input_ids_list], dtype=torch.long)
        attention_mask_tensor = torch.ones_like(input_ids_tensor)

        data = {
            "input_ids": input_ids_tensor,
            "attention_mask": attention_mask_tensor,
        }
        if audio_batch is not None and seqlens_tensor is not None:
            data["audio_data"] = audio_batch
            data["audio_data_seqlens"] = seqlens_tensor
        return BatchFeature(data=data, tensor_type=return_tensors)

    def batch_decode(self, *args, **kwargs):
        return self._base_tokenizer.batch_decode(*args, **kwargs)

    def decode(self, *args, **kwargs):
        return self._base_tokenizer.decode(*args, **kwargs)


__all__ = ["MelConfig", "MossAudioProcessor"]