587 lines
22 KiB
Python
587 lines
22 KiB
Python
from fastapi import FastAPI, UploadFile, HTTPException, Body
|
|
from PIL import Image
|
|
import pytesseract
|
|
from doctr.models import ocr_predictor
|
|
from doctr.io import DocumentFile
|
|
from PyPDF2 import PdfReader
|
|
import camelot
|
|
import spacy
|
|
import logging
|
|
import io
|
|
from logging.handlers import RotatingFileHandler
|
|
import re
|
|
|
|
LOG_PATH = "/var/log/automation-service.log"
|
|
|
|
file_handler = RotatingFileHandler(
|
|
LOG_PATH,
|
|
maxBytes=10*1024*1024,
|
|
backupCount=5,
|
|
encoding="utf-8"
|
|
)
|
|
file_handler.setFormatter(logging.Formatter(
|
|
"%(asctime)s - %(levelname)s - %(name)s - %(message)s"
|
|
))
|
|
|
|
# class AHKParser:
|
|
# lab="AHK"
|
|
# def parse(self,text):
|
|
# r=empty_weight_report("AHK")
|
|
# inv=section(text,"INVOICE WEIGHTS","Bales Weighed")
|
|
# land=section(text,"Bales Weighed","Outturn")
|
|
# loss=section(text,"LOSS","Invoice average")
|
|
|
|
# r["report"]["reference"]=safe_search(r"(AHK\s*/\S+)",text)
|
|
# r["report"]["date"]=extract("Produced On",text)
|
|
|
|
# r["contract"]["invoice_no"]=extract("Client Reference",text)
|
|
# r["contract"]["origin"]=extract("Growth",text)
|
|
# r["contract"]["commodity"]="Raw Cotton"
|
|
|
|
# r["parties"]["seller"]=extract("Client",text)
|
|
# r["parties"]["buyer"]=extract("Buyer",text)
|
|
|
|
# r["shipment"]["vessel"]=extract("Vessel",text)
|
|
# r["shipment"]["bl_no"]=extract("B/L No",text)
|
|
# r["shipment"]["port_destination"]=extract("Destination",text)
|
|
# r["shipment"]["arrival_date"]=extract("Arrival Date",text)
|
|
# r["shipment"]["weighing_method"]=extract("Weighing method",text)
|
|
# r["shipment"]["bales"]=to_float(extract("Total Bales",text))
|
|
|
|
# r["weights"]["gross_landed_kg"]=to_float(extract("Gross",land))
|
|
# r["weights"]["tare_kg"]=to_float(extract("Tare",land))
|
|
# r["weights"]["net_landed_kg"]=to_float(extract("Net",land))
|
|
# r["weights"]["invoice_net_kg"]=to_float(extract("Net",inv))
|
|
# r["weights"]["gain_loss_kg"]=to_float(extract("kg",loss))
|
|
# r["weights"]["gain_loss_percent"]=to_float(extract("Percentage",loss))
|
|
# return r
|
|
class AHKParser:
|
|
lab = "AHK"
|
|
|
|
# ---------- helpers ----------
|
|
def norm(self, s):
|
|
return s.replace("\xa0", " ").strip()
|
|
|
|
def find(self, pattern, text):
|
|
m = re.search(pattern, text, re.I)
|
|
return self.norm(m.group(1)) if m else None
|
|
|
|
def lines(self, text):
|
|
return [self.norm(l) for l in text.splitlines() if self.norm(l)]
|
|
|
|
def block(self, text, labels):
|
|
L = self.lines(text)
|
|
idx = [i for i,l in enumerate(L) if l in labels]
|
|
values = []
|
|
|
|
for i in range(len(idx)):
|
|
start = idx[i]
|
|
end = idx[i+1] if i+1 < len(idx) else len(L)
|
|
for j in range(start, end):
|
|
if L[j].startswith(":"):
|
|
values.append(L[j][1:].strip())
|
|
break
|
|
return dict(zip(labels, values))
|
|
|
|
# ---------- parser ----------
|
|
def parse(self, text):
|
|
r = empty_weight_report("AHK")
|
|
|
|
# ---------- report ----------
|
|
r["report"]["reference"] = self.find(r"(AHK\s*/\S+)", text)
|
|
r["report"]["date"] = self.find(r"Produced On\s*([0-9A-Za-z ]+)", text)
|
|
|
|
# ---------- contract ----------
|
|
r["contract"]["invoice_no"] = self.find(r"Client Reference:\s*([A-Z0-9\- /]+)", text)
|
|
r["contract"]["commodity"] = "Raw Cotton"
|
|
|
|
# ---------- parties ----------
|
|
r["parties"]["buyer"] = self.find(r"Buyer\s*:\s*(.+)", text)
|
|
|
|
# ---------- shipment ----------
|
|
ship = self.block(text, [
|
|
"Total Bales","Vessel","Voy. No.","B/L No.","B/L Date","Destination",
|
|
"Growth","Arrival Date","First date of weighing","Last Date of Weighing",
|
|
"Weighing method","Tare"
|
|
])
|
|
|
|
r["shipment"]["bales"] = to_float(ship.get("Total Bales"))
|
|
r["shipment"]["vessel"] = ship.get("Vessel")
|
|
r["shipment"]["bl_no"] = ship.get("B/L No.")
|
|
r["shipment"]["port_destination"] = ship.get("Destination")
|
|
r["shipment"]["arrival_date"] = ship.get("Arrival Date")
|
|
r["shipment"]["weighing_method"] = ship.get("Weighing method")
|
|
r["contract"]["origin"] = ship.get("Growth")
|
|
|
|
# ---------- weights ----------
|
|
inv = self.block(text, ["Bales","Gross","Tare","Net"])
|
|
land = self.block(section(text,"Bales Weighed","Outturn"),["Bales","Gross","Tare","Net"])
|
|
loss = section(text,"LOSS","Invoice average")
|
|
|
|
r["weights"]["gross_landed_kg"] = to_float(land.get("Gross"))
|
|
r["weights"]["tare_kg"] = to_float(land.get("Tare"))
|
|
r["weights"]["net_landed_kg"] = to_float(land.get("Net"))
|
|
r["weights"]["invoice_net_kg"] = to_float(inv.get("Net"))
|
|
|
|
r["weights"]["gain_loss_kg"] = to_float(self.find(r"(-?\d+\.?\d*)\s*kg", loss))
|
|
r["weights"]["gain_loss_percent"] = to_float(self.find(r"Percentage\s*:\s*(-?\d+\.?\d*)", loss))
|
|
|
|
return r
|
|
|
|
class IntertekParser:
|
|
lab="INTERTEK"
|
|
def parse(self,text):
|
|
r=empty_weight_report("INTERTEK")
|
|
pct=safe_search(r"([0-9.]+)\s*%",text)
|
|
|
|
r["report"]["reference"]=extract("Global Ref",text)
|
|
r["report"]["file_no"]=extract("Report / File No",text)
|
|
r["report"]["date"]=extract("Dated",text)
|
|
|
|
r["contract"]["contract_no"]=extract("Contract No",text)
|
|
r["contract"]["invoice_no"]=extract("Invoice No",text)
|
|
r["contract"]["origin"]=extract("Growth",text)
|
|
r["contract"]["commodity"]="Raw Cotton"
|
|
|
|
r["parties"]["buyer"]=extract("Buyer",text)
|
|
|
|
r["shipment"]["vessel"]=extract("Vessel",text)
|
|
r["shipment"]["bl_no"]=extract("B/L No",text)
|
|
r["shipment"]["arrival_date"]=extract("Arrival Date",text)
|
|
r["shipment"]["weighing_place"]=extract("Weighed at",text)
|
|
r["shipment"]["bales"]=to_float(extract("Invoice Quantity",text))
|
|
|
|
r["weights"]["gross_landed_kg"]=to_float(extract("Gross",text))
|
|
r["weights"]["tare_kg"]=to_float(extract("Invoice Tare",text))
|
|
r["weights"]["net_landed_kg"]=to_float(extract("Landed Weight",text))
|
|
r["weights"]["invoice_net_kg"]=to_float(extract("Invoice Weight",text))
|
|
r["weights"]["gain_loss_kg"]=to_float(extract("Gain",text))
|
|
r["weights"]["gain_loss_percent"]=to_float(pct)
|
|
return r
|
|
|
|
class RobertsonParser:
|
|
lab="ROBERTSON"
|
|
def parse(self,text):
|
|
r=empty_weight_report("ROBERTSON")
|
|
pct=safe_search(r"([0-9.]+)\s*%",text)
|
|
|
|
r["report"]["reference"]=extract("OUR REF",text)
|
|
r["report"]["date"]=extract("DATE",text)
|
|
|
|
r["contract"]["contract_no"]=extract("CONTRACT NO",text)
|
|
r["contract"]["invoice_no"]=extract("INVOICE NO",text)
|
|
r["contract"]["lc_no"]=extract("LIC NO",text)
|
|
r["contract"]["commodity"]="Raw Cotton"
|
|
|
|
r["parties"]["seller"]=extract("SELLER",text)
|
|
r["parties"]["buyer"]=extract("BUYER",text)
|
|
|
|
r["shipment"]["vessel"]=extract("NAME OF VESSEL",text)
|
|
r["shipment"]["port_loading"]=extract("SAILED FROM",text)
|
|
r["shipment"]["port_destination"]=extract("ARRIVED AT",text)
|
|
r["shipment"]["arrival_date"]=extract("DATE OF ARRIVAL",text)
|
|
r["shipment"]["weighing_place"]=extract("PLACE OF CONTROL",text)
|
|
r["shipment"]["bales"]=to_float(extract("CONSIGNMENT",text))
|
|
|
|
r["weights"]["gross_landed_kg"]=to_float(extract("GROSS",text))
|
|
r["weights"]["tare_kg"]=to_float(extract("TARE",text))
|
|
r["weights"]["net_landed_kg"]=to_float(extract("LANDED NET",text))
|
|
r["weights"]["invoice_net_kg"]=to_float(extract("INVOICE NET",text))
|
|
r["weights"]["gain_loss_kg"]=to_float(extract("GAIN",text))
|
|
r["weights"]["gain_loss_percent"]=to_float(pct)
|
|
return r
|
|
|
|
class SGSParser:
|
|
lab="SGS"
|
|
def parse(self,text):
|
|
r=empty_weight_report("SGS")
|
|
r["report"]["reference"]=extract("LANDING REPORT No",text)
|
|
r["report"]["file_no"]=extract("FILE NO.",text)
|
|
r["report"]["date"]=extract("DATE",text)
|
|
|
|
r["contract"]["contract_no"]=extract("CONTRACT NO.",text)
|
|
r["contract"]["invoice_no"]=extract("INVOICE NO.",text)
|
|
r["contract"]["origin"]=extract("ORIGIN",text)
|
|
r["contract"]["commodity"]=extract("PRODUCT",text)
|
|
|
|
r["parties"]["seller"]=extract("Seller",text)
|
|
r["parties"]["buyer"]=extract("Buyer",text)
|
|
r["parties"]["carrier"]=extract("Carrier",text)
|
|
|
|
r["shipment"]["bl_no"]=extract("B/L no.",text)
|
|
r["shipment"]["port_loading"]=extract("Port of loading",text)
|
|
r["shipment"]["port_destination"]=extract("Port of destination",text)
|
|
r["shipment"]["arrival_date"]=extract("Vessel arrival date",text)
|
|
r["shipment"]["weighing_place"]=extract("Place of weighing",text)
|
|
r["shipment"]["weighing_method"]=extract("Weighing mode",text)
|
|
r["shipment"]["bales"]=to_float(extract("Quantity arrived",text))
|
|
|
|
r["weights"]["gross_landed_kg"]=to_float(extract("Gross landed",text))
|
|
r["weights"]["tare_kg"]=to_float(extract("Tare",text))
|
|
r["weights"]["net_landed_kg"]=to_float(extract("Net landed",text))
|
|
r["weights"]["invoice_net_kg"]=to_float(extract("Net invoiced",text))
|
|
r["weights"]["gain_loss_kg"]=to_float(safe_search(r"Gain.*?([0-9.,]+)\s*kgs",text))
|
|
r["weights"]["gain_loss_percent"]=to_float(safe_search(r"Gain\s*\+?\s*([0-9.,]+)\s*%",text))
|
|
return r
|
|
|
|
class PICLParser:
|
|
lab="PICL"
|
|
def parse(self,text):
|
|
r=empty_weight_report("PICL")
|
|
|
|
r["report"]["reference"]=safe_search(r"No[:\s]+([A-Z0-9\-]+)",text)
|
|
r["report"]["date"]=safe_search(r"(Monday|Tuesday|Wednesday|Thursday|Friday|Saturday|Sunday),?\s*([A-Za-z]+\s+[0-9]{1,2},\s*[0-9]{4})",text,group_index=2)
|
|
|
|
r["contract"]["contract_no"]=extract("Contract/Pl No & Date",text)
|
|
r["contract"]["invoice_no"]=extract("Invoice ilo & Date",text)
|
|
r["contract"]["lc_no"]=extract("L/C No & Date",text)
|
|
r["contract"]["origin"]=extract("Country of Origin",text)
|
|
r["contract"]["commodity"]=extract("Commodity",text)
|
|
|
|
r["parties"]["seller"]=extract("FAIRCOT SA",text)
|
|
r["parties"]["buyer"]=extract("M/S.",text)
|
|
r["parties"]["carrier"]=extract("Shipping Agent",text)
|
|
|
|
r["shipment"]["vessel"]=extract("Shipped Per Vessel",text)
|
|
r["shipment"]["bl_no"]=extract("B/L No & Date",text)
|
|
r["shipment"]["port_loading"]=extract("Port of Loading",text)
|
|
r["shipment"]["port_destination"]=extract("Port of Discharge",text)
|
|
r["shipment"]["arrival_date"]=extract("Date of Anival & LDL",text)
|
|
r["shipment"]["weighing_place"]=extract("Place & Date of Weighment",text)
|
|
r["shipment"]["weighing_method"]=extract("Method of Weighment",text)
|
|
r["shipment"]["bales"]=to_float(extract("Grand Total",text))
|
|
|
|
r["weights"]["gross_landed_kg"]=to_float(extract("Total;",text))
|
|
r["weights"]["tare_kg"]=to_float(extract("Tare Weight",text))
|
|
r["weights"]["net_landed_kg"]=to_float(extract("Grand Total",text))
|
|
r["weights"]["invoice_net_kg"]=to_float(extract("Invoice weight",text))
|
|
r["weights"]["gain_loss_kg"]=to_float(safe_search(r"(-[0-9.,]+)\s*KGS",text))
|
|
r["weights"]["gain_loss_percent"]=to_float(safe_search(r"\(\s*([0-9.,]+)\s*o/o\s*\)",text))
|
|
return r
|
|
|
|
|
|
# Configure root logger explicitly
|
|
root = logging.getLogger()
|
|
root.setLevel(logging.INFO)
|
|
root.addHandler(file_handler)
|
|
root.addHandler(logging.StreamHandler())
|
|
|
|
# Use root logger for your app
|
|
logger = logging.getLogger(__name__)
|
|
|
|
app = FastAPI()
|
|
logger.info("Loading models...")
|
|
|
|
nlp = spacy.load("en_core_web_sm")
|
|
predictor = ocr_predictor(pretrained=True)
|
|
|
|
logger.info("Models loaded successfully.")
|
|
|
|
# =============================
|
|
# 🧠 Smart OCR
|
|
# =============================
|
|
@app.post("/ocr")
|
|
async def ocr(file: UploadFile):
|
|
logger.info(f"Received OCR request: {file.filename}")
|
|
try:
|
|
file_data = await file.read()
|
|
ext = file.filename.lower()
|
|
|
|
# --------- PDF with native text ---------
|
|
if ext.endswith(".pdf"):
|
|
logger.info("PDF detected → Extracting native text first")
|
|
reader = PdfReader(io.BytesIO(file_data))
|
|
direct_text = "".join(
|
|
page.extract_text() or "" for page in reader.pages
|
|
)
|
|
|
|
if direct_text.strip():
|
|
logger.info("Native PDF text found → No OCR needed")
|
|
return {"ocr_text": direct_text}
|
|
|
|
# -------- Fallback: scanned PDF OCR --------
|
|
logger.info("No native text → PDF treated as scanned → OCR")
|
|
from pdf2image import convert_from_bytes
|
|
images = convert_from_bytes(file_data)
|
|
text = ""
|
|
for i, img in enumerate(images):
|
|
logger.info(f"OCR page {i+1}/{len(images)}")
|
|
text += pytesseract.image_to_string(img) + "\n"
|
|
|
|
return {"ocr_text": text}
|
|
|
|
# --------- Image file OCR ---------
|
|
logger.info("Image detected → Running OCR")
|
|
img = Image.open(io.BytesIO(file_data))
|
|
text = pytesseract.image_to_string(img)
|
|
return {"ocr_text": text}
|
|
|
|
except Exception as e:
|
|
logger.error(f"OCR failed: {e}", exc_info=True)
|
|
raise HTTPException(status_code=500, detail=str(e))
|
|
|
|
# =============================
|
|
# 🧱 Structure / Layout
|
|
# =============================
|
|
@app.post("/structure")
|
|
async def structure(file: UploadFile):
|
|
logger.info(f"Received structure request: {file.filename}")
|
|
try:
|
|
file_data = await file.read()
|
|
ext = file.filename.lower()
|
|
|
|
if ext.endswith(".pdf"):
|
|
doc = DocumentFile.from_pdf(file_data)
|
|
logger.info(f"Structure prediction on PDF ({len(doc)} pages)")
|
|
else:
|
|
img = Image.open(io.BytesIO(file_data)).convert("RGB")
|
|
doc = DocumentFile.from_images([img])
|
|
logger.info("Structure prediction on image")
|
|
|
|
res = predictor(doc)
|
|
return {"structure": str(res)}
|
|
|
|
except Exception as e:
|
|
logger.error(f"Structure extraction failed: {e}", exc_info=True)
|
|
raise HTTPException(status_code=500, detail=str(e))
|
|
|
|
# =============================
|
|
# 📊 Tables extraction (PDF only)
|
|
# =============================
|
|
@app.post("/tables")
|
|
async def tables(file: UploadFile):
|
|
logger.info(f"Received table extraction request: {file.filename}")
|
|
try:
|
|
file_data = await file.read()
|
|
buffer = io.BytesIO(file_data)
|
|
|
|
tables = camelot.read_pdf(buffer)
|
|
logger.info(f"Found {len(tables)} tables")
|
|
return {"tables": [t.df.to_dict() for t in tables]}
|
|
|
|
except Exception as e:
|
|
logger.error(f"Table extraction failed: {e}", exc_info=True)
|
|
raise HTTPException(status_code=500, detail=str(e))
|
|
|
|
def safe_search(pattern, text, default=None, group_index=1, context=""):
|
|
"""Recherche sécurisée avec logging en cas d'absence de correspondance."""
|
|
m = re.search(pattern, text, re.I | re.S)
|
|
if not m:
|
|
logger.warning("Pattern not found for %s: %s", context, pattern)
|
|
return default
|
|
try:
|
|
return m.group(group_index).strip()
|
|
except IndexError:
|
|
logger.warning("Group index %d not found for %s: %s", group_index, context, pattern)
|
|
return default
|
|
|
|
def to_float(s):
|
|
if not s:
|
|
return None
|
|
s = s.replace(",", "").replace("Kgs", "").replace("kg", "").replace("%", "")
|
|
s = s.replace("lbs", "").replace("LBS", "")
|
|
s = s.strip()
|
|
try:
|
|
return float(s)
|
|
except:
|
|
return None
|
|
|
|
def section(text, start, end=None):
|
|
"""Extract a block of text between two headings, safely."""
|
|
pattern_start = re.escape(start)
|
|
if end:
|
|
pattern_end = re.escape(end)
|
|
reg = re.compile(pattern_start + r"(.*?)" + pattern_end, re.S | re.I)
|
|
else:
|
|
reg = re.compile(pattern_start + r"(.*)", re.S | re.I)
|
|
m = reg.search(text)
|
|
if not m:
|
|
logger.warning("Section not found: start='%s', end='%s'", start, end)
|
|
return ""
|
|
return m.group(1).strip()
|
|
|
|
def extract_field(text, label, default=None):
|
|
"""Extract a line of the form 'Label: value', safely."""
|
|
pattern = rf"{re.escape(label)}\s*:?[\s]+([^\n]+)"
|
|
return safe_search(pattern, text, default=default, context=f"field '{label}'")
|
|
|
|
def extract(label, text, default=None):
|
|
"""
|
|
Robust extraction for OCR/PDF text.
|
|
Works with:
|
|
Label: Value
|
|
Label Value
|
|
Label .... Value
|
|
"""
|
|
if not text:
|
|
return default
|
|
|
|
patterns = [
|
|
rf"{re.escape(label)}\s*[:\-]?\s*([^\n\r]+)",
|
|
rf"{re.escape(label)}\s+([^\n\r]+)"
|
|
]
|
|
|
|
for p in patterns:
|
|
m = re.search(p, text, re.I)
|
|
if m:
|
|
return m.group(1).strip()
|
|
|
|
return default
|
|
|
|
def extract_report_metadata(text):
|
|
logger.info("Starting metadata extraction, text length=%d", len(text))
|
|
|
|
try:
|
|
# ----------- SECTIONS -----------
|
|
order_details = section(text, "Order details", "Weights")
|
|
invoice_section = section(text, "INVOICE WEIGHTS", "Bales Weighed")
|
|
landed_section = section(text, "Bales Weighed", "Outturn")
|
|
loss_section = section(text, "LOSS", "Invoice average")
|
|
avg_section = section(text, "Invoice average", "Comments")
|
|
signature_block = section(text, "Signed on")
|
|
|
|
# ----------- TOP INFO -----------
|
|
top_info = {
|
|
"produced_on": extract_field(text, "Produced On"),
|
|
"printed_date": extract_field(text, "Printed Date"),
|
|
"client_reference": extract_field(text, "Client Reference"),
|
|
"report_number": safe_search(r"(AHK\S+)", text, default="", context="report_number", group_index=1),
|
|
}
|
|
|
|
# ----------- ORDER DETAILS -----------
|
|
parties = {
|
|
"client": extract_field(order_details, "Client"),
|
|
"client_ref_no": extract_field(order_details, "Client Ref No"),
|
|
"buyer": extract_field(order_details, "Buyer"),
|
|
"destination": extract_field(order_details, "Destination"),
|
|
}
|
|
|
|
shipment = {
|
|
"total_bales": extract_field(order_details, "Total Bales"),
|
|
"vessel": extract_field(order_details, "Vessel"),
|
|
"voyage_no": extract_field(order_details, "Voy. No"),
|
|
"bl_no": extract_field(order_details, "B/L No"),
|
|
"bl_date": extract_field(order_details, "B/L Date"),
|
|
"growth": extract_field(order_details, "Growth"),
|
|
"arrival_date": extract_field(order_details, "Arrival Date"),
|
|
"first_weighing_date": extract_field(order_details, "First date of weighing"),
|
|
"last_weighing_date": extract_field(order_details, "Last Date of Weighing"),
|
|
"weighing_method": extract_field(order_details, "Weighing method"),
|
|
"tare_basis": extract_field(order_details, "Tare"),
|
|
}
|
|
|
|
# ----------- INVOICE SECTION -----------
|
|
invoice = {
|
|
"bales": extract_field(invoice_section, "Bales"),
|
|
"gross": extract_field(invoice_section, "Gross"),
|
|
"tare": extract_field(invoice_section, "Tare"),
|
|
"net": extract_field(invoice_section, "Net"),
|
|
}
|
|
|
|
# ----------- LANDED SECTION -----------
|
|
landed = {
|
|
"bales": extract_field(landed_section, "Bales"),
|
|
"gross": extract_field(landed_section, "Gross"),
|
|
"tare": extract_field(landed_section, "Tare"),
|
|
"net": extract_field(landed_section, "Net"),
|
|
}
|
|
|
|
# ----------- LOSS SECTION -----------
|
|
loss = {
|
|
"kg": extract_field(loss_section, "kg"),
|
|
"lb": extract_field(loss_section, "lb"),
|
|
"percent": extract_field(loss_section, "Percentage"),
|
|
}
|
|
|
|
# ----------- AVERAGES SECTION -----------
|
|
averages = {
|
|
"invoice_gross_per_bale": extract_field(avg_section, "Invoice average"),
|
|
"landed_gross_per_bale": extract_field(avg_section, "Landed average"),
|
|
}
|
|
|
|
# ----------- SIGNATURE -----------
|
|
signature = {
|
|
"signed_on": extract_field(signature_block, "Signed on"),
|
|
"signed_by": safe_search(r"\n([A-Za-z ]+)\nClient Services", signature_block, default="", context="signed_by"),
|
|
"role": "Client Services Coordinator",
|
|
"company": "Alfred H. Knight International Limited"
|
|
}
|
|
|
|
logger.info("Metadata extraction completed successfully")
|
|
return {
|
|
"report": top_info,
|
|
"parties": parties,
|
|
"shipment": shipment,
|
|
"weights": {
|
|
"invoice": invoice,
|
|
"landed": landed,
|
|
"loss": loss,
|
|
"averages": averages
|
|
},
|
|
"signature": signature
|
|
}
|
|
|
|
except Exception as e:
|
|
logger.exception("Unexpected error during metadata extraction")
|
|
raise HTTPException(status_code=500, detail=f"Metadata extraction failed: {e}")
|
|
|
|
def detect_template(text):
|
|
t = text.lower()
|
|
|
|
if "alfred h. knight" in t and "cotton landing report" in t:
|
|
return "AHK"
|
|
|
|
if "intertek" in t and "landing report" in t:
|
|
return "INTERTEK"
|
|
|
|
if "robertson international" in t or "ri ref no" in t:
|
|
return "ROBERTSON"
|
|
|
|
if "landing report" in t and "carcon cargo" in t:
|
|
return "SGS"
|
|
|
|
if "pacific inspection company" in t or "picl-bd.com" in t:
|
|
return "PICL"
|
|
|
|
return "UNKNOWN"
|
|
|
|
@app.post("/metadata")
|
|
async def metadata(text: str = Body(..., embed=True)):
|
|
return extract_report_metadata(text)
|
|
|
|
@app.post("/parse")
|
|
async def parse_endpoint(text: str = Body(..., embed=True)):
|
|
return parse_report(text)
|
|
|
|
PARSERS = {
|
|
"AHK": AHKParser(),
|
|
"INTERTEK": IntertekParser(),
|
|
"ROBERTSON": RobertsonParser(),
|
|
"SGS": SGSParser(),
|
|
"PICL": PICLParser()
|
|
}
|
|
|
|
def empty_weight_report(lab):
|
|
return {
|
|
"lab": lab,
|
|
"report": {"reference": None, "file_no": None, "date": None},
|
|
"contract": {"contract_no": None, "invoice_no": None, "lc_no": None, "origin": None, "commodity": None},
|
|
"parties": {"seller": None, "buyer": None, "carrier": None},
|
|
"shipment": {
|
|
"vessel": None, "bl_no": None, "port_loading": None,
|
|
"port_destination": None, "arrival_date": None,
|
|
"weighing_place": None, "weighing_method": None,
|
|
"bales": None
|
|
},
|
|
"weights": {
|
|
"gross_landed_kg": None, "tare_kg": None,
|
|
"net_landed_kg": None, "invoice_net_kg": None,
|
|
"gain_loss_kg": None, "gain_loss_percent": None
|
|
}
|
|
}
|
|
|
|
def parse_report(text):
|
|
template=detect_template(text)
|
|
if template not in PARSERS:
|
|
return {"template":"UNKNOWN"}
|
|
return PARSERS[template].parse(text) |