Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -14,7 +14,9 @@ from transformers import (
|
|
| 14 |
)
|
| 15 |
from peft import PeftModel
|
| 16 |
|
| 17 |
-
#
|
|
|
|
|
|
|
| 18 |
_SPACE_VARIANTS = r"[\u202f\u00a0\u2009\u200a\u2060]"
|
| 19 |
|
| 20 |
def _normalise_apostrophes(text: str) -> str:
|
|
@@ -31,40 +33,62 @@ def _clean_timex(ent: str) -> str:
|
|
| 31 |
ent = ent.replace("</s>", "").strip()
|
| 32 |
return re.sub(r"[\.]+$", "", ent)
|
| 33 |
|
| 34 |
-
#
|
|
|
|
|
|
|
| 35 |
NER_ID = "Rhulli/Roberta-ner-temporal-expresions-secondtrain"
|
| 36 |
ID2LABEL = {0: "O", 1: "B-TIMEX", 2: "I-TIMEX"}
|
| 37 |
BASE_ID = "google/gemma-2b-it"
|
| 38 |
ADAPTER_ID = "Rhulli/gemma-2b-it-TIMEX3"
|
| 39 |
|
| 40 |
-
#
|
|
|
|
|
|
|
| 41 |
quant_config = BitsAndBytesConfig(
|
| 42 |
load_in_4bit=True,
|
| 43 |
bnb_4bit_quant_type="nf4",
|
| 44 |
bnb_4bit_compute_dtype=torch.float16,
|
| 45 |
)
|
| 46 |
|
| 47 |
-
#
|
|
|
|
|
|
|
| 48 |
HF_TOKEN = os.getenv("HF_TOKEN")
|
| 49 |
|
|
|
|
|
|
|
|
|
|
| 50 |
def load_models():
|
|
|
|
| 51 |
ner_tok = AutoTokenizer.from_pretrained(NER_ID, token=HF_TOKEN)
|
| 52 |
ner_mod = AutoModelForTokenClassification.from_pretrained(NER_ID, token=HF_TOKEN)
|
| 53 |
ner_mod.eval()
|
| 54 |
if torch.cuda.is_available():
|
| 55 |
ner_mod.to("cuda")
|
| 56 |
|
|
|
|
| 57 |
base_mod = AutoModelForCausalLM.from_pretrained(
|
| 58 |
BASE_ID,
|
| 59 |
-
|
| 60 |
-
|
|
|
|
|
|
|
|
|
|
| 61 |
)
|
| 62 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
norm_mod = PeftModel.from_pretrained(
|
| 64 |
base_mod,
|
| 65 |
ADAPTER_ID,
|
| 66 |
-
|
| 67 |
-
|
|
|
|
| 68 |
)
|
| 69 |
norm_mod.eval()
|
| 70 |
|
|
@@ -72,9 +96,18 @@ def load_models():
|
|
| 72 |
|
| 73 |
# Carga inicial de los modelos
|
| 74 |
ner_tok, ner_mod, norm_tok, norm_mod = load_models()
|
| 75 |
-
eos_id = norm_tok.convert_tokens_to_ids("<end_of_turn>")
|
| 76 |
|
| 77 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 78 |
def read_file(file_obj) -> str:
|
| 79 |
path = file_obj.name
|
| 80 |
if path.lower().endswith('.pdf'):
|
|
@@ -93,7 +126,9 @@ def read_file(file_obj) -> str:
|
|
| 93 |
except:
|
| 94 |
return data.decode('latin-1', errors='ignore')
|
| 95 |
|
| 96 |
-
#
|
|
|
|
|
|
|
| 97 |
def extract_timex(text: str):
|
| 98 |
text_norm = _normalise_spaces(_normalise_apostrophes(text))
|
| 99 |
inputs = ner_tok(text_norm, return_tensors="pt", truncation=True)
|
|
@@ -124,6 +159,9 @@ def extract_timex(text: str):
|
|
| 124 |
|
| 125 |
return [_clean_timex(e) for e in entities]
|
| 126 |
|
|
|
|
|
|
|
|
|
|
| 127 |
def normalize_timex(expr: str, dct: str) -> str:
|
| 128 |
prompt = (
|
| 129 |
f"<start_of_turn>user\n"
|
|
@@ -132,8 +170,15 @@ def normalize_timex(expr: str, dct: str) -> str:
|
|
| 132 |
f"Expresi贸n Original: {expr}<end_of_turn>\n"
|
| 133 |
f"<start_of_turn>model\n"
|
| 134 |
)
|
| 135 |
-
|
| 136 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 137 |
|
| 138 |
full_decoded = norm_tok.decode(
|
| 139 |
outputs[0, inputs.input_ids.shape[1]:],
|
|
@@ -142,11 +187,14 @@ def normalize_timex(expr: str, dct: str) -> str:
|
|
| 142 |
raw_tag = full_decoded.split("<end_of_turn>")[0].strip()
|
| 143 |
return raw_tag.replace("[", "<").replace("]", ">")
|
| 144 |
|
| 145 |
-
#
|
|
|
|
|
|
|
| 146 |
def run_pipeline(files, raw_text, dct):
|
| 147 |
rows = []
|
| 148 |
file_list = files if isinstance(files, list) else ([files] if files else [])
|
| 149 |
|
|
|
|
| 150 |
if raw_text:
|
| 151 |
for line in raw_text.splitlines():
|
| 152 |
if line.strip():
|
|
@@ -156,6 +204,7 @@ def run_pipeline(files, raw_text, dct):
|
|
| 156 |
'Normalizaci贸n': normalize_timex(expr, dct)
|
| 157 |
})
|
| 158 |
|
|
|
|
| 159 |
for f in file_list:
|
| 160 |
content = read_file(f)
|
| 161 |
for line in content.splitlines():
|
|
@@ -172,29 +221,30 @@ def run_pipeline(files, raw_text, dct):
|
|
| 172 |
|
| 173 |
return df, ""
|
| 174 |
|
| 175 |
-
#
|
|
|
|
|
|
|
| 176 |
with gr.Blocks() as demo:
|
| 177 |
-
gr.Markdown(
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
| 182 |
-
|
| 183 |
-
|
| 184 |
-
|
| 185 |
-
|
| 186 |
-
|
| 187 |
-
|
| 188 |
-
|
| 189 |
-
|
| 190 |
-
|
| 191 |
-
|
| 192 |
-
|
| 193 |
-
)
|
| 194 |
|
| 195 |
with gr.Row():
|
| 196 |
with gr.Column(scale=1):
|
| 197 |
-
files = gr.File(file_types=['.txt'], file_count='multiple', label='Archivos (.txt)')
|
| 198 |
dct_input = gr.Textbox(value="2025-06-11", label="Fecha de Anclaje (YYYY-MM-DD)")
|
| 199 |
run_btn = gr.Button("Procesar")
|
| 200 |
with gr.Column(scale=2):
|
|
@@ -203,29 +253,28 @@ with gr.Blocks() as demo:
|
|
| 203 |
output_table = gr.Dataframe(headers=['Expresi贸n', 'Normalizaci贸n'], label="Resultados", type="pandas")
|
| 204 |
output_logs = gr.Textbox(label="Logs", lines=5, interactive=False)
|
| 205 |
|
| 206 |
-
# Despu茅s de definir output_table y output_logs:
|
| 207 |
download_btn = gr.Button("Descargar CSV")
|
| 208 |
-
csv_file_output
|
| 209 |
|
| 210 |
-
#
|
| 211 |
run_btn.click(
|
| 212 |
fn=run_pipeline,
|
| 213 |
inputs=[files, raw_text, dct_input],
|
| 214 |
outputs=[output_table, output_logs]
|
| 215 |
)
|
| 216 |
|
| 217 |
-
#
|
| 218 |
def export_csv(df):
|
| 219 |
csv_path = "resultados.csv"
|
| 220 |
df.to_csv(csv_path, index=False)
|
| 221 |
return gr.update(value=csv_path, visible=True)
|
| 222 |
|
| 223 |
-
# Asociar el bot贸n de descarga al CSV
|
| 224 |
download_btn.click(
|
| 225 |
fn=export_csv,
|
| 226 |
inputs=[output_table],
|
| 227 |
outputs=[csv_file_output]
|
| 228 |
)
|
| 229 |
|
| 230 |
-
|
| 231 |
-
|
|
|
|
|
|
| 14 |
)
|
| 15 |
from peft import PeftModel
|
| 16 |
|
| 17 |
+
# =========================
|
| 18 |
+
# Utilidades de normalizaci贸n
|
| 19 |
+
# =========================
|
| 20 |
_SPACE_VARIANTS = r"[\u202f\u00a0\u2009\u200a\u2060]"
|
| 21 |
|
| 22 |
def _normalise_apostrophes(text: str) -> str:
|
|
|
|
| 33 |
ent = ent.replace("</s>", "").strip()
|
| 34 |
return re.sub(r"[\.]+$", "", ent)
|
| 35 |
|
| 36 |
+
# =========================
|
| 37 |
+
# Identificadores de modelos
|
| 38 |
+
# =========================
|
| 39 |
NER_ID = "Rhulli/Roberta-ner-temporal-expresions-secondtrain"
|
| 40 |
ID2LABEL = {0: "O", 1: "B-TIMEX", 2: "I-TIMEX"}
|
| 41 |
BASE_ID = "google/gemma-2b-it"
|
| 42 |
ADAPTER_ID = "Rhulli/gemma-2b-it-TIMEX3"
|
| 43 |
|
| 44 |
+
# =========================
|
| 45 |
+
# Cuantizaci贸n 4-bit (NF4)
|
| 46 |
+
# =========================
|
| 47 |
quant_config = BitsAndBytesConfig(
|
| 48 |
load_in_4bit=True,
|
| 49 |
bnb_4bit_quant_type="nf4",
|
| 50 |
bnb_4bit_compute_dtype=torch.float16,
|
| 51 |
)
|
| 52 |
|
| 53 |
+
# =========================
|
| 54 |
+
# Token de HF (si lo usas privado)
|
| 55 |
+
# =========================
|
| 56 |
HF_TOKEN = os.getenv("HF_TOKEN")
|
| 57 |
|
| 58 |
+
# =========================
|
| 59 |
+
# Carga de modelos
|
| 60 |
+
# =========================
|
| 61 |
def load_models():
|
| 62 |
+
# --- NER ---
|
| 63 |
ner_tok = AutoTokenizer.from_pretrained(NER_ID, token=HF_TOKEN)
|
| 64 |
ner_mod = AutoModelForTokenClassification.from_pretrained(NER_ID, token=HF_TOKEN)
|
| 65 |
ner_mod.eval()
|
| 66 |
if torch.cuda.is_available():
|
| 67 |
ner_mod.to("cuda")
|
| 68 |
|
| 69 |
+
# --- Base Causal LM (Gemma 2B-it) con 4-bit ---
|
| 70 |
base_mod = AutoModelForCausalLM.from_pretrained(
|
| 71 |
BASE_ID,
|
| 72 |
+
token=HF_TOKEN,
|
| 73 |
+
device_map="auto", # deja a Accelerate decidir
|
| 74 |
+
quantization_config=quant_config, # aplica 4-bit NF4
|
| 75 |
+
torch_dtype=torch.float16,
|
| 76 |
+
low_cpu_mem_usage=True,
|
| 77 |
)
|
| 78 |
+
|
| 79 |
+
# --- Tokenizer del BASE (no del adapter) ---
|
| 80 |
+
norm_tok = AutoTokenizer.from_pretrained(BASE_ID, use_fast=True, token=HF_TOKEN)
|
| 81 |
+
# Asegurar pad_token si falta
|
| 82 |
+
if norm_tok.pad_token is None and norm_tok.eos_token is not None:
|
| 83 |
+
norm_tok.pad_token = norm_tok.eos_token
|
| 84 |
+
|
| 85 |
+
# --- Inyectar el LoRA SIN device_map (evitar meta/offload issues) ---
|
| 86 |
norm_mod = PeftModel.from_pretrained(
|
| 87 |
base_mod,
|
| 88 |
ADAPTER_ID,
|
| 89 |
+
token=HF_TOKEN,
|
| 90 |
+
is_trainable=False,
|
| 91 |
+
offload_state_dict=False,
|
| 92 |
)
|
| 93 |
norm_mod.eval()
|
| 94 |
|
|
|
|
| 96 |
|
| 97 |
# Carga inicial de los modelos
|
| 98 |
ner_tok, ner_mod, norm_tok, norm_mod = load_models()
|
|
|
|
| 99 |
|
| 100 |
+
# Determinar eos_id de manera segura
|
| 101 |
+
try:
|
| 102 |
+
eos_id = norm_tok.convert_tokens_to_ids("<end_of_turn>")
|
| 103 |
+
if eos_id is None or eos_id == norm_tok.unk_token_id:
|
| 104 |
+
eos_id = norm_tok.eos_token_id
|
| 105 |
+
except Exception:
|
| 106 |
+
eos_id = norm_tok.eos_token_id
|
| 107 |
+
|
| 108 |
+
# =========================
|
| 109 |
+
# Lectura de archivos (.txt, .pdf)
|
| 110 |
+
# =========================
|
| 111 |
def read_file(file_obj) -> str:
|
| 112 |
path = file_obj.name
|
| 113 |
if path.lower().endswith('.pdf'):
|
|
|
|
| 126 |
except:
|
| 127 |
return data.decode('latin-1', errors='ignore')
|
| 128 |
|
| 129 |
+
# =========================
|
| 130 |
+
# Extracci贸n NER de TIMEX
|
| 131 |
+
# =========================
|
| 132 |
def extract_timex(text: str):
|
| 133 |
text_norm = _normalise_spaces(_normalise_apostrophes(text))
|
| 134 |
inputs = ner_tok(text_norm, return_tensors="pt", truncation=True)
|
|
|
|
| 159 |
|
| 160 |
return [_clean_timex(e) for e in entities]
|
| 161 |
|
| 162 |
+
# =========================
|
| 163 |
+
# Normalizaci贸n con Gemma + LoRA
|
| 164 |
+
# =========================
|
| 165 |
def normalize_timex(expr: str, dct: str) -> str:
|
| 166 |
prompt = (
|
| 167 |
f"<start_of_turn>user\n"
|
|
|
|
| 170 |
f"Expresi贸n Original: {expr}<end_of_turn>\n"
|
| 171 |
f"<start_of_turn>model\n"
|
| 172 |
)
|
| 173 |
+
device = next(norm_mod.parameters()).device
|
| 174 |
+
inputs = norm_tok(prompt, return_tensors="pt").to(device)
|
| 175 |
+
with torch.no_grad():
|
| 176 |
+
outputs = norm_mod.generate(
|
| 177 |
+
**inputs,
|
| 178 |
+
max_new_tokens=64,
|
| 179 |
+
eos_token_id=eos_id,
|
| 180 |
+
do_sample=False,
|
| 181 |
+
)
|
| 182 |
|
| 183 |
full_decoded = norm_tok.decode(
|
| 184 |
outputs[0, inputs.input_ids.shape[1]:],
|
|
|
|
| 187 |
raw_tag = full_decoded.split("<end_of_turn>")[0].strip()
|
| 188 |
return raw_tag.replace("[", "<").replace("]", ">")
|
| 189 |
|
| 190 |
+
# =========================
|
| 191 |
+
# Pipeline principal
|
| 192 |
+
# =========================
|
| 193 |
def run_pipeline(files, raw_text, dct):
|
| 194 |
rows = []
|
| 195 |
file_list = files if isinstance(files, list) else ([files] if files else [])
|
| 196 |
|
| 197 |
+
# Texto pegado
|
| 198 |
if raw_text:
|
| 199 |
for line in raw_text.splitlines():
|
| 200 |
if line.strip():
|
|
|
|
| 204 |
'Normalizaci贸n': normalize_timex(expr, dct)
|
| 205 |
})
|
| 206 |
|
| 207 |
+
# Archivos subidos
|
| 208 |
for f in file_list:
|
| 209 |
content = read_file(f)
|
| 210 |
for line in content.splitlines():
|
|
|
|
| 221 |
|
| 222 |
return df, ""
|
| 223 |
|
| 224 |
+
# =========================
|
| 225 |
+
# Interfaz Gradio
|
| 226 |
+
# =========================
|
| 227 |
with gr.Blocks() as demo:
|
| 228 |
+
gr.Markdown("""
|
| 229 |
+
## TIMEX Extractor & Normalizer
|
| 230 |
+
|
| 231 |
+
Esta aplicaci贸n permite **extraer** expresiones temporales de textos o archivos (.txt, .pdf)
|
| 232 |
+
y **normalizarlas** a formato **TIMEX3**.
|
| 233 |
+
|
| 234 |
+
**C贸mo usar:**
|
| 235 |
+
1. Sube uno o varios archivos en la columna izquierda.
|
| 236 |
+
2. Ajusta la *Fecha de Anclaje (DCT)*.
|
| 237 |
+
3. Escribe o pega tu texto en la columna derecha.
|
| 238 |
+
4. Pulsa **Procesar** para ver los resultados.
|
| 239 |
+
|
| 240 |
+
**Columnas de salida:**
|
| 241 |
+
- **Expresi贸n**: la frase temporal extra铆da.
|
| 242 |
+
- **Normalizaci贸n**: la etiqueta TIMEX3 generada.
|
| 243 |
+
""")
|
|
|
|
| 244 |
|
| 245 |
with gr.Row():
|
| 246 |
with gr.Column(scale=1):
|
| 247 |
+
files = gr.File(file_types=['.txt', '.pdf'], file_count='multiple', label='Archivos (.txt, .pdf)')
|
| 248 |
dct_input = gr.Textbox(value="2025-06-11", label="Fecha de Anclaje (YYYY-MM-DD)")
|
| 249 |
run_btn = gr.Button("Procesar")
|
| 250 |
with gr.Column(scale=2):
|
|
|
|
| 253 |
output_table = gr.Dataframe(headers=['Expresi贸n', 'Normalizaci贸n'], label="Resultados", type="pandas")
|
| 254 |
output_logs = gr.Textbox(label="Logs", lines=5, interactive=False)
|
| 255 |
|
|
|
|
| 256 |
download_btn = gr.Button("Descargar CSV")
|
| 257 |
+
csv_file_output = gr.File(label="Descargar resultados en CSV", visible=False)
|
| 258 |
|
| 259 |
+
# Acci贸n principal de procesamiento
|
| 260 |
run_btn.click(
|
| 261 |
fn=run_pipeline,
|
| 262 |
inputs=[files, raw_text, dct_input],
|
| 263 |
outputs=[output_table, output_logs]
|
| 264 |
)
|
| 265 |
|
| 266 |
+
# Exportar a CSV
|
| 267 |
def export_csv(df):
|
| 268 |
csv_path = "resultados.csv"
|
| 269 |
df.to_csv(csv_path, index=False)
|
| 270 |
return gr.update(value=csv_path, visible=True)
|
| 271 |
|
|
|
|
| 272 |
download_btn.click(
|
| 273 |
fn=export_csv,
|
| 274 |
inputs=[output_table],
|
| 275 |
outputs=[csv_file_output]
|
| 276 |
)
|
| 277 |
|
| 278 |
+
# Lanzar la app (Spaces recoger谩 host/port)
|
| 279 |
+
if __name__ == "__main__":
|
| 280 |
+
demo.launch()
|