backend added

Dockerfile

@@ -0,0 +1,11 @@
+FROM python:3.10-slim
+
+WORKDIR /code
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY . .
+
+EXPOSE 7860
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

app.py

@@ -0,0 +1,160 @@
+from pydantic import BaseModel
+from fastapi import FastAPI, HTTPException, UploadFile, File
+from fastapi.middleware.cors import CORSMiddleware
+from pydantic import Field
+import torch
+from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassification, AutoConfig
+import torch.nn.functional as F
+import fitz
+
+# -----------------------------------------
+# GLOBAL PDF CACHE
+# -----------------------------------------
+pdf_cache = {"text": None}
+
+# -----------------------------------------
+# HUGGINGFACE MODEL PATHS
+# -----------------------------------------
+SUMMARY_MODEL = "krrishsinha/legal_summariser"
+QNA_MODEL = "krrishsinha/nlpques-ans"
+CLAUSE_MODEL = "krrishsinha/clausedetectionfinal"
+
+
+# -----------------------------------------
+# PDF READER
+# -----------------------------------------
+def pdfopen(filepath: str) -> str:
+    doc = fitz.open(filepath)
+    text = ""
+    for page in doc:
+        text += page.get_text()
+    doc.close()
+    return text.strip()
+
+
+# -----------------------------------------
+# SUMMARIZER PIPELINE
+# -----------------------------------------
+def summarizer():
+    return pipeline("summarization", model=SUMMARY_MODEL)
+
+
+# -----------------------------------------
+# QNA PIPELINE
+# -----------------------------------------
+def anq():
+    return pipeline("question-answering", model=QNA_MODEL)
+
+
+# -----------------------------------------
+# CLAUSE DETECTION
+# -----------------------------------------
+def clause(sen):
+    
+    tokenizer = AutoTokenizer.from_pretrained(CLAUSE_MODEL)
+    model = AutoModelForSequenceClassification.from_pretrained(CLAUSE_MODEL)
+    config = AutoConfig.from_pretrained(CLAUSE_MODEL)
+
+    inputs = tokenizer(sen, return_tensors="pt", truncation=True, padding=True)
+
+    with torch.no_grad():
+        outputs = model(**inputs)
+        logits = outputs.logits
+        pred_id = int(torch.argmax(logits, dim=1).item())
+
+    predicted_label = config.id2label.get(pred_id, f"LABEL_{pred_id}")
+    return predicted_label
+
+
+# -----------------------------------------
+# FASTAPI APP
+# -----------------------------------------
+app = FastAPI()
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+
+@app.get("/")
+def welcome():
+    return {"welcome": "Lawlytics AI Corporate Legal Intelligence"}
+
+
+# -----------------------------------------
+# PDF UPLOAD
+# -----------------------------------------
+@app.post("/upload")
+async def uploading(file: UploadFile = File(...)):
+    try:
+        file_path = f"./{file.filename}"
+        with open(file_path, "wb") as f:
+            f.write(await file.read())
+
+        t = pdfopen(file_path)
+
+        if not t:
+            raise HTTPException(status_code=400, detail="No text found in PDF")
+
+        pdf_cache["text"] = t
+        return {"message": "PDF processed successfully", "characters_extracted": len(t)}
+
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+
+# -----------------------------------------
+# SUMMARISATION
+# -----------------------------------------
+@app.post("/summarise")
+def summary():
+    txt = pdf_cache["text"]
+    if not txt:
+        raise HTTPException(status_code=400, detail="Upload PDF first")
+
+    summarise_fn = summarizer()
+    output = summarise_fn(txt, max_length=100, min_length=30, do_sample=False)
+
+    return {"summary": output}
+
+
+# -----------------------------------------
+# QUESTION ANSWERING
+# -----------------------------------------
+class QnaRequest(BaseModel):
+    question: str
+    context: str = None
+
+
+@app.post("/qna")
+def quesans(payload: QnaRequest):
+    if not pdf_cache["text"] and not payload.context:
+        raise HTTPException(status_code=400, detail="Upload PDF first")
+
+    context = payload.context or pdf_cache["text"]
+
+    qna_fn = anq()
+    result = qna_fn(question=payload.question, context=context)
+
+    return {"answer": result["answer"]}
+
+
+# -----------------------------------------
+# CLAUSE DETECTION
+# -----------------------------------------
+class ClauseRequest(BaseModel):
+    text: str = None
+
+
+@app.post("/clausedetection")
+def clausing(payload: ClauseRequest):
+    text = payload.text or pdf_cache["text"]
+    if not text:
+        raise HTTPException(status_code=400, detail="Provide text or upload PDF first")
+
+    detected = clause(text)
+    return {"detected_clause": detected}

requirements.txt

@@ -0,0 +1,9 @@
+fastapi
+uvicorn
+python-multipart
+pymupdf
+transformers
+torch
+huggingface_hub
+safetensors
+numpy