Spaces:
Sleeping
Sleeping
import re | |
from fastapi import FastAPI, HTTPException | |
from fastapi.responses import StreamingResponse | |
from fastapi.responses import JSONResponse | |
from pydantic import BaseModel | |
from huggingface_hub import InferenceClient | |
import uvicorn | |
from typing import Generator, List | |
import json # Asegúrate de que esta línea esté al principio del archivo | |
import nltk | |
import os | |
import google.protobuf # This line should execute without errors if protobuf is installed correctly | |
import sentencepiece | |
from transformers import pipeline, AutoTokenizer,AutoModelForSequenceClassification,AutoModel | |
import spacy | |
import numpy as np | |
import torch | |
nltk.data.path.append(os.getenv('NLTK_DATA')) | |
app = FastAPI() | |
# Initialize the InferenceClient with your model | |
client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1") | |
class Item(BaseModel): | |
prompt: str | |
history: list | |
system_prompt: str | |
temperature: float = 0.8 | |
max_new_tokens: int = 4000 | |
top_p: float = 0.15 | |
repetition_penalty: float = 1.0 | |
def format_prompt(current_prompt, history): | |
formatted_history = "<s>" | |
for entry in history: | |
if entry["role"] == "user": | |
formatted_history += f"[USER] {entry['content']} [/USER]" | |
elif entry["role"] == "assistant": | |
formatted_history += f"[ASSISTANT] {entry['content']} [/ASSISTANT]" | |
formatted_history += f"[USER] {current_prompt} [/USER]</s>" | |
return formatted_history | |
def generate_stream(item: Item) -> Generator[bytes, None, None]: | |
formatted_prompt = format_prompt(f"{item.system_prompt}, {item.prompt}", item.history) | |
# Estimate token count for the formatted_prompt | |
input_token_count = len(nltk.word_tokenize(formatted_prompt)) # NLTK tokenization | |
# Ensure total token count doesn't exceed the maximum limit | |
max_tokens_allowed = 32768 | |
max_new_tokens_adjusted = max(1, min(item.max_new_tokens, max_tokens_allowed - input_token_count)) | |
generate_kwargs = { | |
"temperature": item.temperature, | |
"max_new_tokens": max_new_tokens_adjusted, | |
"top_p": item.top_p, | |
"repetition_penalty": item.repetition_penalty, | |
"do_sample": True, | |
"seed": 42, | |
} | |
# Stream the response from the InferenceClient | |
for response in client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True): | |
# This assumes 'details=True' gives you a structure where you can access the text like this | |
chunk = { | |
"text": response.token.text, | |
"complete": response.generated_text is not None # Adjust based on how you detect completion | |
} | |
yield json.dumps(chunk).encode("utf-8") + b"\n" | |
class SummarizeRequest(BaseModel): | |
text: str | |
async def generate_text(item: Item): | |
# Stream response back to the client | |
return StreamingResponse(generate_stream(item), media_type="application/x-ndjson") | |
# Define request model | |
class TextRequest(BaseModel): | |
text: str # Single string of long text | |
# Load Longformer model and tokenizer | |
tokenizer = AutoTokenizer.from_pretrained("allenai/longformer-base-4096") | |
model = AutoModel.from_pretrained("allenai/longformer-base-4096") | |
# Endpoint to process the document and return embeddings | |
async def process_document(request: TextRequest): | |
try: | |
# Split the text into segments that fit within the model's max input size | |
max_length = 4096 # Maximum token length for Longformer | |
words = request.text.split() | |
tokens = tokenizer.encode(request.text, add_special_tokens=True) | |
input_ids = [] | |
current_chunk = [] | |
for token in tokens: | |
if len(current_chunk) + len(tokenizer.convert_ids_to_tokens([token])) < max_length: | |
current_chunk.append(token) | |
else: | |
input_ids.append(current_chunk) | |
current_chunk = [token] | |
if current_chunk: | |
input_ids.append(current_chunk) # Add the last chunk if any | |
# Generate embeddings for each segment | |
embeddings_list = [] | |
for ids in input_ids: | |
inputs = {'input_ids': torch.tensor(ids).unsqueeze(0)} # Batch size 1 | |
outputs = model(**inputs) | |
embeddings = outputs.last_hidden_state.mean(dim=1).detach().numpy() | |
embeddings_list.append(embeddings.tolist()) # Store embeddings for each segment | |
return { | |
"embeddings": embeddings_list | |
} | |
except Exception as e: | |
print(f"Error during document processing: {e}") | |
raise HTTPException(status_code=500, detail=str(e)) | |
# @app.post("/summarize") | |
# async def summarize(request: TextRequest): | |
# try: | |
# # Preprocess and segment the text | |
# processed_text = preprocess_text(request.text) | |
# segments = segment_text(processed_text) | |
# # Classify each segment safely | |
# classified_segments = [] | |
# for segment in segments: | |
# try: | |
# result = classifier(segment) | |
# classified_segments.append(result) | |
# except Exception as e: | |
# print(f"Error classifying segment: {e}") | |
# classified_segments.append({"error": str(e)}) | |
# # Optional: Reduce tokens or summarize | |
# reduced_texts = [] | |
# for segment in segments: | |
# try: | |
# reduced_text, token_count = reduce_tokens(segment) | |
# reduced_texts.append((reduced_text, token_count)) | |
# except Exception as e: | |
# print(f"Error during token reduction: {e}") | |
# reduced_texts.append(("Error", 0)) | |
# return { | |
# "classified_segments": classified_segments, | |
# "reduced_texts": reduced_texts | |
# } | |
# except Exception as e: | |
# print(f"Error during token reduction: {e}") | |
# raise HTTPException(status_code=500, detail=str(e)) | |
# if __name__ == "__main__": | |
# uvicorn.run(app, host="0.0.0.0", port=8000) | |