app.py

import sys
import os
import json
import shutil
import re
import gc
import time
from datetime import datetime
from typing import List, Tuple, Dict, Union
import pandas as pd
import pdfplumber
import gradio as gr
import torch
import matplotlib.pyplot as plt
from fpdf import FPDF
import unicodedata

# === Configuration ===
persistent_dir = "/data/hf_cache"
model_cache_dir = os.path.join(persistent_dir, "txagent_models")
tool_cache_dir = os.path.join(persistent_dir, "tool_cache")
file_cache_dir = os.path.join(persistent_dir, "cache")
report_dir = os.path.join(persistent_dir, "reports")

for d in [model_cache_dir, tool_cache_dir, file_cache_dir, report_dir]:
    os.makedirs(d, exist_ok=True)

os.environ["HF_HOME"] = model_cache_dir
os.environ["TRANSFORMERS_CACHE"] = model_cache_dir

current_dir = os.path.dirname(os.path.abspath(__file__))
src_path = os.path.abspath(os.path.join(current_dir, "src"))
sys.path.insert(0, src_path)

from txagent.txagent import TxAgent

MAX_MODEL_TOKENS = 131072
MAX_NEW_TOKENS = 4096
MAX_CHUNK_TOKENS = 8192
BATCH_SIZE = 1
PROMPT_OVERHEAD = 300
SAFE_SLEEP = 0.5

def estimate_tokens(text: str) -> int:
    return len(text) // 4 + 1

def clean_response(text: str) -> str:
    text = re.sub(r"\[.*?\]|\bNone\b", "", text, flags=re.DOTALL)
    text = re.sub(r"\n{3,}", "\n\n", text)
    return text.strip()

def remove_duplicate_paragraphs(text: str) -> str:
    paragraphs = text.strip().split("\n\n")
    seen = set()
    unique_paragraphs = []
    for p in paragraphs:
        clean_p = p.strip()
        if clean_p and clean_p not in seen:
            unique_paragraphs.append(clean_p)
            seen.add(clean_p)
    return "\n\n".join(unique_paragraphs)

# === FastAPI for mobile API endpoint ===
from fastapi import FastAPI, UploadFile, File
from fastapi.responses import JSONResponse
import uvicorn

app = FastAPI()

@app.post("/analyze")
async def analyze_file_api(file: UploadFile = File(...)):
    agent = init_agent()
    temp_file_path = os.path.join(file_cache_dir, file.filename)
    with open(temp_file_path, "wb") as f:
        f.write(await file.read())
    messages = []
    messages, pdf_path = process_report(agent, open(temp_file_path, "rb"), messages)
    if pdf_path:
        return JSONResponse(content={"summary": messages[-2]['content'], "pdf": pdf_path})
    return JSONResponse(content={"error": "Processing failed."}, status_code=400)

# === Original Gradio UI launch preserved ===
if __name__ == "__main__":
    agent = init_agent()
    ui = create_ui(agent)
    import threading
    threading.Thread(target=lambda: ui.launch(server_name="0.0.0.0", server_port=7860, allowed_paths=["/data/hf_cache/reports"], share=False)).start()
    uvicorn.run(app, host="0.0.0.0", port=8000)