Update app.py

app.py

@@ -1,52 +1,105 @@
+import os, tempfile, uuid, pdfplumber
 import gradio as gr
-import os
 from groq import Groq
+from docx import Document                # python-docx
 
-# Initialize Groq client
-groq_api_key = "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"
-client = Groq(api_key=groq_api_key)
-
-lang_options = ["English", "Urdu", "Spanish", "French", "Arabic", "Chinese", "German"]
-
-def translate_with_groq(text, src_lang, tgt_lang):
-    if src_lang == tgt_lang:
-        return text
+# ── Groq setup ──────────────────────────────────────────────────────────────────
+client = Groq("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")
+MODEL_NAME = "mixtral-8x7b"
+LANGS      = ["English", "Urdu", "Spanish", "French", "Arabic", "Chinese", "German","Hindi","Turkish"]
 
+# ── Core translator (single chunk) ──────────────────────────────────────────────
+def llm_translate(chunk: str, src: str, tgt: str) -> str:
+    if src == tgt:          # shortcut
+        return chunk
     prompt = (
-        f"You are a translation assistant. Translate the following text from {src_lang} to {tgt_lang}:\n\n\"{text}\""
+        "You are a professional multilingual translator. "
+        f"Translate the following text from {src} to {tgt}.\n\n{chunk}"
+    )
+    resp = client.chat.completions.create(
+        model=MODEL_NAME,
+        messages=[{"role": "system", "content": "Translator"}, {"role": "user", "content": prompt}],
+        temperature=0.2,
+        max_completion_tokens=2048,
     )
+    return resp.choices[0].message.content.strip()
+
+# ── Document helpers ────────────────────────────────────────────────────────────
+MAX_CHARS = 6000                     # stay far below LLM context; tune as needed
+
+def translate_long_text(text, src, tgt):
+    """Chunk long text on paragraph boundaries and re‑assemble."""
+    out, buff = [], []
+    for para in text.splitlines(keepends=True):
+        buff.append(para)
+        if sum(len(p) for p in buff) >= MAX_CHARS:
+            out.append(llm_translate("".join(buff), src, tgt))
+            buff = []
+    if buff:
+        out.append(llm_translate("".join(buff), src, tgt))
+    return "\n".join(out)
 
-    try:
-        resp = client.chat.completions.create(
-            model="mistral-saba-24b",
-            messages=[
-                {"role": "system", "content": "You are a multilingual translation assistant."},
-                {"role": "user", "content": prompt}
-            ],
-            temperature=0.2,
-            max_completion_tokens=512
-        )
-        return resp.choices[0].message.content.strip()
-    except Exception as e:
-        return f"❌ Error: {e}"
-
-# Gradio UI (Blocks with improved layout)
+def handle_docx(path, src, tgt):
+    doc_in  = Document(path)
+    doc_out = Document()
+    for para in doc_in.paragraphs:
+        translated = translate_long_text(para.text, src, tgt)
+        doc_out.add_paragraph(translated)
+    out_path = os.path.join(tempfile.gettempdir(), f"{uuid.uuid4()}.docx")
+    doc_out.save(out_path)
+    return out_path
+
+def handle_pdf(path, src, tgt):
+    txt = []
+    with pdfplumber.open(path) as pdf:
+        for page in pdf.pages:
+            txt.append(page.extract_text() or "")
+    translated_text = translate_long_text("\n".join(txt), src, tgt)
+    # save as .txt for simplicity; could generate PDF if you prefer
+    out_path = os.path.join(tempfile.gettempdir(), f"{uuid.uuid4()}.txt")
+    with open(out_path, "w", encoding="utf-8") as f:
+        f.write(translated_text)
+    return out_path
+
+# ── Gradio UI (Blocks) ──────────────────────────────────────────────────────────
 with gr.Blocks(theme=gr.themes.Default(primary_hue="teal", font="poppins")) as demo:
-    gr.Markdown("# 🌍 AI Multilingual Translator")
-    gr.Markdown("Translate between multiple languages using **Groq's Mistral-Saba 24B** model.")
+    gr.Markdown("# 🌍 AI Multilingual Translator\nBlazing‑fast Groq (Mistral‑Saba 24B)")
+
+    with gr.Tab("✏️ Text"):
+        with gr.Row():
+            with gr.Column():
+                txt_in  = gr.Textbox(lines=5, label="Input")
+                src1    = gr.Dropdown(LANGS, value="English", label="From")
+                tgt1    = gr.Dropdown(LANGS, value="Urdu",     label="To")
+                btn1    = gr.Button("🔁 Translate", variant="primary")
+            with gr.Column():
+                txt_out = gr.Textbox(lines=5, label="Translation")
+        btn1.click(llm_translate, inputs=[txt_in, src1, tgt1], outputs=txt_out)
+
+    with gr.Tab("📄 Document"):
+        with gr.Row():
+            with gr.Column():
+                file_in = gr.File(label="Upload PDF or DOCX")
+                src2    = gr.Dropdown(LANGS, value="English", label="From")
+                tgt2    = gr.Dropdown(LANGS, value="Urdu",     label="To")
+                btn2    = gr.Button("🚀 Translate File", variant="primary")
+            with gr.Column():
+                file_out = gr.File(label="Download translated file")
 
-    with gr.Row():
-        with gr.Column():
-            input_text = gr.Textbox(label="📝 Input Text", lines=5, placeholder="Type your text here...")
-            src_lang = gr.Dropdown(lang_options, value="English", label="🌐 From")
-            tgt_lang = gr.Dropdown(lang_options, value="Urdu", label="🌐 To")
-            translate_btn = gr.Button("🔁 Translate", variant="primary")
-        with gr.Column():
-            output_text = gr.Textbox(label="✅ Translated Text", lines=5)
+        def translate_file(file, src, tgt):
+            if file is None:
+                raise gr.Error("Please upload a file.")
+            ext = os.path.splitext(file.name)[1].lower()
+            if ext == ".docx":
+                return handle_docx(file.name, src, tgt)
+            elif ext == ".pdf":
+                return handle_pdf(file.name, src, tgt)
+            else:
+                raise gr.Error("Unsupported format (only PDF & DOCX).")
 
-    gr.Markdown("---\nPowered by Groq & Gradio")
+        btn2.click(translate_file, inputs=[file_in, src2, tgt2], outputs=file_out)
 
-    translate_btn.click(translate_with_groq, inputs=[input_text, src_lang, tgt_lang], outputs=output_text)
+    gr.Markdown("---\nMade with ❤️ by Groq & Gradio")
 
 if __name__ == "__main__":
     demo.launch()