Gitead
5cdd3452f2
Claude vision's self-reported text coordinates had a small but consistent bias, causing the patient-row double-click to land above the name and the Exam-column divider count to occasionally miscount and hard-fail. Swap patient-row, Exam, and Save button targeting to exact OCR bounding boxes (Google Vision via the existing PaymentOCRService) instead of AI-estimated ratios.
234 lines
7.9 KiB
Python
Executable File
234 lines
7.9 KiB
Python
Executable File
from fastapi import FastAPI, UploadFile, File, HTTPException, Request
|
|
from fastapi.middleware.cors import CORSMiddleware
|
|
from fastapi.responses import StreamingResponse, JSONResponse, PlainTextResponse
|
|
from typing import List, Optional
|
|
import io
|
|
import os
|
|
import asyncio
|
|
import tempfile
|
|
import uvicorn
|
|
from dotenv import load_dotenv
|
|
load_dotenv()
|
|
|
|
from complete_pipeline_adapter import process_images_to_rows,rows_to_csv_bytes
|
|
from complete_pipeline import extract_words_and_text
|
|
from pdf_extractor import extract_ra_pdf
|
|
|
|
app = FastAPI(
|
|
title="Payment OCR Services API",
|
|
description="FastAPI wrapper around the OCR pipeline (Google Vision + deskew + line grouping + extraction).",
|
|
version="1.0.0",
|
|
)
|
|
|
|
# Concurrency/semaphore (optional but useful for OCR)
|
|
MAX_CONCURRENCY = int(os.getenv("MAX_CONCURRENCY", "2"))
|
|
semaphore = asyncio.Semaphore(MAX_CONCURRENCY)
|
|
|
|
active_jobs = 0
|
|
waiting_jobs = 0
|
|
lock = asyncio.Lock()
|
|
|
|
# CORS
|
|
cors_origins = os.getenv("CORS_ORIGINS", "*")
|
|
allow_origins = ["*"] if cors_origins.strip() == "*" else [o.strip() for o in cors_origins.split(",") if o.strip()]
|
|
app.add_middleware(
|
|
CORSMiddleware,
|
|
allow_origins=allow_origins,
|
|
allow_methods=["*"],
|
|
allow_headers=["*"],
|
|
)
|
|
|
|
ALLOWED_EXTS = {".jpg", ".jpeg", ".png", ".tif", ".tiff", ".bmp"}
|
|
ALLOWED_PDF_EXTS = {".pdf"}
|
|
|
|
# -------------------------------------------------
|
|
# Health + status
|
|
# -------------------------------------------------
|
|
@app.get("/health", response_class=PlainTextResponse)
|
|
def health():
|
|
creds = os.environ.get("GOOGLE_APPLICATION_CREDENTIALS", "")
|
|
return f"OK | GOOGLE_APPLICATION_CREDENTIALS set: {bool(creds)}"
|
|
|
|
@app.get("/status")
|
|
async def get_status():
|
|
async with lock:
|
|
return {
|
|
"active_jobs": active_jobs,
|
|
"queued_jobs": waiting_jobs,
|
|
"max_concurrency": MAX_CONCURRENCY,
|
|
"status": "busy" if active_jobs > 0 or waiting_jobs > 0 else "idle",
|
|
}
|
|
|
|
# -------------------------------------------------
|
|
# Helpers
|
|
# -------------------------------------------------
|
|
def _validate_files(files: List[UploadFile]):
|
|
if not files:
|
|
raise HTTPException(status_code=400, detail="No files provided.")
|
|
bad = [f.filename for f in files if os.path.splitext(f.filename or "")[1].lower() not in ALLOWED_EXTS]
|
|
if bad:
|
|
raise HTTPException(
|
|
status_code=415,
|
|
detail=f"Unsupported file types: {', '.join(bad)}. Allowed: {', '.join(sorted(ALLOWED_EXTS))}"
|
|
)
|
|
|
|
# -------------------------------------------------
|
|
# Endpoints
|
|
# -------------------------------------------------
|
|
@app.post("/extract/json")
|
|
async def extract_json(files: List[UploadFile] = File(...)):
|
|
_validate_files(files)
|
|
|
|
async with lock:
|
|
global waiting_jobs
|
|
waiting_jobs += 1
|
|
|
|
async with semaphore:
|
|
async with lock:
|
|
waiting_jobs -= 1
|
|
global active_jobs
|
|
active_jobs += 1
|
|
|
|
try:
|
|
blobs = [await f.read() for f in files]
|
|
names = [f.filename or "upload.bin" for f in files]
|
|
rows = process_images_to_rows(blobs, names) # calls pipeline
|
|
return JSONResponse(content={"rows": rows})
|
|
except Exception as e:
|
|
raise HTTPException(status_code=500, detail=f"Processing error: {e}")
|
|
finally:
|
|
async with lock:
|
|
active_jobs -= 1
|
|
|
|
@app.post("/extract/words")
|
|
async def extract_words(file: UploadFile = File(...)):
|
|
"""Raw OCR only — no deskew/line-grouping/payment-domain logic, just the
|
|
flat word list with pixel bounding boxes. Used by the Windows Type Agent
|
|
to locate exact click targets (e.g. a patient name) instead of relying
|
|
on an AI-vision estimate of where text is."""
|
|
_validate_files([file])
|
|
|
|
async with lock:
|
|
global waiting_jobs
|
|
waiting_jobs += 1
|
|
|
|
async with semaphore:
|
|
async with lock:
|
|
waiting_jobs -= 1
|
|
global active_jobs
|
|
active_jobs += 1
|
|
|
|
tmp_path = None
|
|
try:
|
|
suffix = os.path.splitext(file.filename or "")[1] or ".png"
|
|
blob = await file.read()
|
|
with tempfile.NamedTemporaryFile(delete=False, suffix=suffix) as tmp:
|
|
tmp.write(blob)
|
|
tmp_path = tmp.name
|
|
words, _ = extract_words_and_text(tmp_path)
|
|
return JSONResponse(content={"words": words})
|
|
except Exception as e:
|
|
raise HTTPException(status_code=500, detail=f"OCR error: {e}")
|
|
finally:
|
|
if tmp_path and os.path.exists(tmp_path):
|
|
os.remove(tmp_path)
|
|
async with lock:
|
|
active_jobs -= 1
|
|
|
|
@app.post("/extract/csvtext", response_class=PlainTextResponse)
|
|
async def extract_csvtext(files: List[UploadFile] = File(...)):
|
|
_validate_files(files)
|
|
|
|
async with lock:
|
|
global waiting_jobs
|
|
waiting_jobs += 1
|
|
|
|
async with semaphore:
|
|
async with lock:
|
|
waiting_jobs -= 1
|
|
global active_jobs
|
|
active_jobs += 1
|
|
|
|
try:
|
|
blobs = [await f.read() for f in files]
|
|
names = [f.filename or "upload.bin" for f in files]
|
|
rows = process_images_to_rows(blobs, names)
|
|
csv_bytes = rows_to_csv_bytes(rows)
|
|
return PlainTextResponse(csv_bytes.decode("utf-8"), media_type="text/csv")
|
|
except Exception as e:
|
|
raise HTTPException(status_code=500, detail=f"Processing error: {e}")
|
|
finally:
|
|
async with lock:
|
|
active_jobs -= 1
|
|
|
|
@app.post("/extract/pdf/json")
|
|
async def extract_pdf_json(files: List[UploadFile] = File(...)):
|
|
bad = [f.filename for f in files if os.path.splitext(f.filename or "")[1].lower() not in ALLOWED_PDF_EXTS]
|
|
if bad:
|
|
raise HTTPException(status_code=415, detail=f"Only PDF files allowed. Got: {', '.join(bad)}")
|
|
|
|
async with lock:
|
|
global waiting_jobs
|
|
waiting_jobs += 1
|
|
|
|
async with semaphore:
|
|
async with lock:
|
|
waiting_jobs -= 1
|
|
global active_jobs
|
|
active_jobs += 1
|
|
|
|
try:
|
|
all_rows = []
|
|
all_headers = []
|
|
for f in files:
|
|
blob = await f.read()
|
|
result = extract_ra_pdf(blob, f.filename or "upload.pdf")
|
|
all_rows.extend(result["rows"])
|
|
all_headers.append(result["header"])
|
|
return JSONResponse(content={"rows": all_rows, "headers": all_headers})
|
|
except Exception as e:
|
|
raise HTTPException(status_code=500, detail=f"PDF extraction error: {e}")
|
|
finally:
|
|
async with lock:
|
|
active_jobs -= 1
|
|
|
|
|
|
@app.post("/extract/csv")
|
|
async def extract_csv(files: List[UploadFile] = File(...), filename: Optional[str] = None):
|
|
_validate_files(files)
|
|
|
|
async with lock:
|
|
global waiting_jobs
|
|
waiting_jobs += 1
|
|
|
|
async with semaphore:
|
|
async with lock:
|
|
waiting_jobs -= 1
|
|
global active_jobs
|
|
active_jobs += 1
|
|
|
|
try:
|
|
blobs = [await f.read() for f in files]
|
|
names = [f.filename or "upload.bin" for f in files]
|
|
rows = process_images_to_rows(blobs, names)
|
|
csv_bytes = rows_to_csv_bytes(rows)
|
|
out_name = filename or "medical_billing_extract.csv"
|
|
return StreamingResponse(
|
|
io.BytesIO(csv_bytes),
|
|
media_type="text/csv",
|
|
headers={"Content-Disposition": f'attachment; filename="{out_name}"'}
|
|
)
|
|
except Exception as e:
|
|
raise HTTPException(status_code=500, detail=f"Processing error: {e}")
|
|
finally:
|
|
async with lock:
|
|
active_jobs -= 1
|
|
|
|
# -------------------------------------------------
|
|
# Entrypoint (same pattern as your selenium app)
|
|
# -------------------------------------------------
|
|
if __name__ == "__main__":
|
|
host = os.getenv("HOST")
|
|
port = int(os.getenv("PORT"))
|
|
uvicorn.run(app, host=host, port=port)
|