plan-genie-ai / main.py
Yassine
Fix the allowed entities for each type
7b036e8
from fastapi import FastAPI, Body
import torch
import spacy
import os
from pathlib import Path
from fastapi.middleware.cors import CORSMiddleware
from transformers import AutoTokenizer, AutoModelForTokenClassification, AutoModelForSequenceClassification
from pydantic import BaseModel
# Define input model
class TextInput(BaseModel):
text: str
# Initialize FastAPI
app = FastAPI()
# Add CORS middleware
app.add_middleware(
CORSMiddleware,
# Vous pouvez restreindre ceci à votre frontend spécifique
allow_origins=["*"],
allow_credentials=True,
allow_methods=["*"],
allow_headers=["*"],
)
# Get base directory
base_dir = Path(__file__).parent.absolute()
# Your Hugging Face Hub username
HF_USERNAME = "YassineJedidi" # Replace with your actual username
# Définition des entités valides pour chaque type
entites_valides = {
"Tâche": {"TITRE", "DELAI", "PRIORITE"},
"Événement": {"TITRE", "DATE_HEURE"},
}
# Try to load models from Hugging Face Hub
try:
print("Loading models from Hugging Face Hub")
# Model repositories on Hugging Face
tokenizer_repo = f"{HF_USERNAME}/tasks-tokenizer"
ner_model_repo = f"{HF_USERNAME}/tasks-ner"
type_model_repo = f"{HF_USERNAME}/tasks-type"
print(f"Loading tokenizer from: {tokenizer_repo}")
print(f"Loading NER model from: {ner_model_repo}")
print(f"Loading type model from: {type_model_repo}")
# Load models from Hugging Face Hub
tokenizer = AutoTokenizer.from_pretrained(tokenizer_repo)
ner_model = AutoModelForTokenClassification.from_pretrained(ner_model_repo)
type_model = AutoModelForSequenceClassification.from_pretrained(
type_model_repo)
except Exception as e:
print(f"Error loading models from Hugging Face Hub: {e}")
# Fallback to local files if available
try:
# Convert paths to strings with forward slashes
tokenizer_path = str(base_dir / "models" /
"tasks-tokenizer").replace("\\", "/")
ner_model_path = str(base_dir / "models" /
"tasks-ner").replace("\\", "/")
type_model_path = str(base_dir / "models" /
"tasks-types").replace("\\", "/")
print(f"Falling back to local models")
print(f"Loading tokenizer from: {tokenizer_path}")
print(f"Loading NER model from: {ner_model_path}")
print(f"Loading type model from: {type_model_path}")
# Load models from local files
tokenizer = AutoTokenizer.from_pretrained(
tokenizer_path, local_files_only=True)
ner_model = AutoModelForTokenClassification.from_pretrained(
ner_model_path, local_files_only=True)
type_model = AutoModelForSequenceClassification.from_pretrained(
type_model_path, local_files_only=True)
except Exception as e:
print(f"Error loading local models: {e}")
# Fallback to base model from HuggingFace
print("Falling back to base CamemBERT model from HuggingFace Hub")
tokenizer = AutoTokenizer.from_pretrained("camembert-base")
ner_model = AutoModelForTokenClassification.from_pretrained(
"camembert-base")
type_model = AutoModelForSequenceClassification.from_pretrained(
"camembert-base")
# Load spaCy for tokenization
nlp = spacy.load('fr_core_news_lg')
# Set device (CPU or GPU)
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
ner_model = ner_model.to(device)
type_model = type_model.to(device)
# Retrieve label mappings
id2label = ner_model.config.id2label
id2type = type_model.config.id2label
@app.get("/")
def root():
return {"message": "FastAPI NLP Model is running!"}
@app.post("/predict-type/")
async def predict_type(input_data: TextInput):
text = input_data.text
inputs = tokenizer(text, return_tensors="pt",
truncation=True, padding=True).to(device)
with torch.no_grad():
outputs = type_model(**inputs)
predicted_class_id = outputs.logits.argmax().item()
predicted_type = id2type[predicted_class_id]
confidence = torch.softmax(outputs.logits, dim=1).max().item()
return {"type": predicted_type, "confidence": confidence}
@app.post("/extract-entities/")
async def extract_entities(input_data: TextInput):
text = input_data.text
doc = nlp(text)
tokens = [token.text for token in doc]
inputs = tokenizer(tokens, is_split_into_words=True,
return_tensors="pt", truncation=True, padding=True).to(device)
with torch.no_grad():
outputs = ner_model(**inputs)
predictions = outputs.logits.argmax(dim=2)
entities = {}
current_entity = None
current_text = []
word_ids = inputs.word_ids(0)
for idx, word_idx in enumerate(word_ids):
if word_idx is None:
continue
if idx > 0 and word_ids[idx-1] == word_idx:
continue
prediction = predictions[0, idx].item()
predicted_label = id2label[prediction]
if predicted_label.startswith("B-"):
if current_entity:
entity_type = current_entity[2:]
if entity_type not in entities:
entities[entity_type] = []
entities[entity_type].append(" ".join(current_text))
current_entity = predicted_label
current_text = [tokens[word_idx]]
elif predicted_label.startswith("I-") and current_entity and predicted_label[2:] == current_entity[2:]:
current_text.append(tokens[word_idx])
else:
if current_entity:
entity_type = current_entity[2:]
if entity_type not in entities:
entities[entity_type] = []
entities[entity_type].append(" ".join(current_text))
current_entity = None
current_text = []
if current_entity:
entity_type = current_entity[2:]
if entity_type not in entities:
entities[entity_type] = []
entities[entity_type].append(" ".join(current_text))
return {"entities": entities}
@app.post("/analyze-text/")
async def analyze_text(input_data: TextInput):
type_result = await predict_type(input_data)
text_type = type_result["type"]
confidence = type_result["confidence"]
raw_entities = (await extract_entities(input_data))["entities"]
# Filtrage des entités selon le type détecté
allowed = entites_valides.get(text_type, set())
filtered_entities = {k: v for k, v in raw_entities.items() if k in allowed}
return {
"type": text_type,
"confidence": confidence,
"entities": filtered_entities
}