Upload

app.py

@@ -0,0 +1,138 @@
+import gradio as gr
+import requests
+import uuid
+import os
+import pdfplumber
+import docx2txt
+import logging
+import shutil
+from langdetect import detect
+from multilingual_pdf2text.pdf2text import PDF2Text
+from multilingual_pdf2text.models.document_model.document import Document
+
+logging.basicConfig(level=logging.INFO)
+
+FLOWISE_URL = "http://localhost:3000/api/v1/prediction/3d1eebaf-31ee-44d6-94bc-4457948f5240"
+COUNTRIES = ["USA", "Germany", "UAE", "Turkey", "Jordan"]
+LANGUAGES = ["Arabic", "English", "German"]
+
+# Tesseract availability check
+print("🔍 Checking if Tesseract OCR is installed...")
+tess_path = shutil.which("tesseract")
+if tess_path:
+    print(f"✅ Tesseract found at: {tess_path}")
+else:
+    print("❌ Tesseract not found. Arabic PDF extraction will likely fail.")
+
+def extract_text_from_file(file_path):
+    ext = os.path.splitext(file_path)[1].lower()
+
+    try:
+        if ext == ".pdf":
+            with pdfplumber.open(file_path) as pdf:
+                extracted = "\n".join(page.extract_text() for page in pdf.pages if page.extract_text())
+
+            sample = extracted[:1000].strip()
+            if sample:
+                try:
+                    lang = detect(sample)
+                except Exception:
+                    lang = "unknown"
+
+                if lang == "ar":
+                    print("📄 Using Arabic-aware extractor for PDF.")
+                    try:
+                        arabic_doc = Document(document_path=file_path, language="ara")
+                        pdf2text = PDF2Text(document=arabic_doc)
+                        extracted = pdf2text.extract()
+                        if isinstance(extracted, list):
+                            if isinstance(extracted[0], dict):
+                                return "\n".join([item.get("text", "") for item in extracted])
+                            elif isinstance(extracted[0], str):
+                                return "\n".join(extracted)
+                        elif isinstance(extracted, str):
+                            return extracted
+                        else:
+                            return "[Error: Unexpected format in extracted text.]"
+                    except Exception as e:
+                        print(f"❌ Arabic PDF extraction failed: {e}")
+                        return "[Error: Failed to extract Arabic text from PDF.]"
+                else:
+                    return extracted
+            else:
+                arabic_doc = Document(document_path=file_path, language="ara")
+                pdf2text = PDF2Text(document=arabic_doc)
+                extracted = pdf2text.extract()
+                if isinstance(extracted, list):
+                    if isinstance(extracted[0], dict):
+                        return "\n".join([item.get("text", "") for item in extracted])
+                    elif isinstance(extracted[0], str):
+                        return "\n".join(extracted)
+                elif isinstance(extracted, str):
+                    return extracted
+                else:
+                    return "[Error: Unexpected format in extracted text.]"
+
+        elif ext == ".docx":
+            return docx2txt.process(file_path)
+
+        elif ext == ".txt":
+            with open(file_path, "r", encoding="utf-8", errors="ignore") as f:
+                return f.read()
+
+        else:
+            return "[Unsupported file format]"
+
+    except Exception as e:
+        return f"[Error reading file: {e}]"
+
+
+def send_to_flowise(file, country, language):
+    if file is None or not country or not language:
+        return "❗Please upload a contract, select a country and language."
+
+    try:
+        contract_text = extract_text_from_file(file.name)
+        if not contract_text.strip():
+            return "⚠️ Could not read any content from the uploaded file."
+
+        prompt = f"""Please audit the following employment contract in light of {country} labor law.
+Only refer to the provided law document (do not use external knowledge).
+Is the contract compliant? Provide reasoning and highlight any issues, and suggest improvements.
+Be precise and don't miss any details in the contract.
+Please respond in {language}.
+<CONTRACT START>
+{contract_text}
+<CONTRACT END>
+"""
+
+        payload = {
+            "question": prompt,
+            "chatId": str(uuid.uuid4())  
+        }
+
+        res = requests.post(FLOWISE_URL, json=payload, timeout=300)
+        print("🔍 Flowise raw response:", res.text)
+
+        if res.status_code != 200:
+            return f"❌ Flowise error {res.status_code}: {res.text}"
+
+        return res.json().get("text", "⚠️ Flowise returned no text.")
+
+    except Exception as e:
+        return f"❌ Connection error: {e}"
+
+with gr.Blocks(title="Contract Auditor") as demo:
+    gr.Markdown("## 📑 Contract Auditor")
+
+    with gr.Row():
+        country_dropdown = gr.Dropdown(choices=COUNTRIES, label="🌍 Select Country", interactive=True)
+        language_dropdown = gr.Dropdown(choices=LANGUAGES, label="🗣️ Response Language", interactive=True)
+        file_input = gr.File(label="📎 Upload a contract file", file_types=[".pdf", ".docx", ".txt"], type="filepath")
+
+    submit_btn = gr.Button("📤 Audit Contract")
+    output = gr.Textbox(label="🧠 Audit Result", lines=12)
+
+    submit_btn.click(fn=send_to_flowise, inputs=[file_input, country_dropdown, language_dropdown], outputs=output)
+
+demo.launch()