Upload 5 files

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+DejaVuSans.ttf filter=lfs diff=lfs merge=lfs -text

DejaVuSans.ttf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7da195a74c55bef988d0d48f9508bd5d849425c1770dba5d7bfc6ce9ed848954
+size 757076

README.md

@@ -1,13 +1,32 @@
 ---
-title: Audiototext
-emoji: ⚡
-colorFrom: pink
-colorTo: gray
+title: YouTube Transcript & Summary
+emoji: 🎧
+colorFrom: blue
+colorTo: green
 sdk: gradio
 sdk_version: 5.31.0
 app_file: app.py
 pinned: false
-short_description: mfilterit
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# YouTube Transcript, Translation & Summary (Whisper + Hugging Face)
+
+This Space extracts audio from a YouTube video, detects language, transcribes speech using OpenAI Whisper, translates to English if needed, and provides a summary.
+
+**How to use:**
+1. Paste a YouTube URL in the box.
+2. Click "Process".
+3. View detected language, full transcript, English translation, and summary.
+
+**Tech stack:** Gradio, Hugging Face Transformers, OpenAI Whisper, Facebook BART, yt-dlp
+
+---
+
+## Requirements
+- All dependencies listed in requirements.txt.
+- yt-dlp is included as a pip dependency.
+
+---
+
+## Author
+Built for quick demo and prototyping by Jagan (template by ChatGPT).

app.py

@@ -0,0 +1,180 @@
+import gradio as gr
+import openai
+from langdetect import detect
+from transformers import pipeline
+from keybert import KeyBERT
+from fpdf import FPDF
+import os
+import re
+import unicodedata
+
+# --- SETUP ---
+openai.api_key = os.getenv("OPENAI_API_KEY")  # Set in HF Space Secrets
+
+summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
+kw_model = KeyBERT()
+FONT_PATH = "DejaVuSans.ttf"  # Must be uploaded to Space root!
+
+BRANDS = [
+    "Apple", "Google", "Microsoft", "Amazon", "Coca-Cola", "Pepsi", "Samsung", "Nike", "ICICI",
+    "Meta", "Facebook", "Instagram", "YouTube", "Netflix", "Reliance", "Tata", "Airtel", "Jio",
+    "Motilal", "Wipro", "Paytm", "Zomato", "Swiggy", "OLA", "Uber"
+]
+
+def extract_brands(text):
+    found = [brand for brand in BRANDS if brand.lower() in text.lower()]
+    return found if found else ["None detected"]
+
+def extract_topics(text, top_n=5):
+    keywords = kw_model.extract_keywords(text, top_n=top_n, stop_words='english')
+    topics = [kw for kw, score in keywords]
+    return topics if topics else ["None extracted"]
+
+def make_bullets(summary):
+    sentences = summary.replace("\n", " ").split('. ')
+    bullets = [f"- {s.strip()}" for s in sentences if s.strip()]
+    return "\n".join(bullets)
+
+def make_str(val):
+    try:
+        if val is None:
+            return ""
+        if isinstance(val, (bool, int, float)):
+            return str(val)
+        if isinstance(val, list):
+            return "\n".join([make_str(v) for v in val])
+        if isinstance(val, dict):
+            return str(val)
+        return str(val)
+    except Exception:
+        return ""
+
+def very_safe_multicell(pdf, text, w=0, h=8, maxlen=50):
+    """Force-break lines so no line/word exceeds maxlen chars, avoiding fpdf2 crash."""
+    if not isinstance(text, str):
+        text = str(text)
+    # Remove unprintable chars (e.g. control characters)
+    text = "".join(ch for ch in text if unicodedata.category(ch)[0] != "C")
+    # Step 1: break long words
+    def break_long_words(t):
+        lines = []
+        for paragraph in t.split('\n'):
+            for word in paragraph.split(' '):
+                while len(word) > maxlen:
+                    lines.append(word[:maxlen])
+                    word = word[maxlen:]
+                lines.append(word)
+            lines.append('')
+        return '\n'.join(lines)
+    text = break_long_words(text)
+    # Step 2: ensure no line is too long (wrap at maxlen)
+    wrapped = []
+    for line in text.splitlines():
+        while len(line) > maxlen:
+            wrapped.append(line[:maxlen])
+            line = line[maxlen:]
+        wrapped.append(line)
+    safe_text = '\n'.join(wrapped)
+    pdf.multi_cell(w, h, safe_text)
+
+def create_pdf_report(language, transcript_en, brands, topics, key_takeaways):
+    pdf = FPDF()
+    pdf.set_auto_page_break(auto=True, margin=10)
+    pdf.set_margins(left=10, top=10, right=10)
+    pdf.add_font("DejaVu", style="", fname=FONT_PATH, uni=True)
+    pdf.add_font("DejaVu", style="B", fname=FONT_PATH, uni=True)
+    pdf.add_page()
+    pdf.set_font("DejaVu", "B", 16)
+    pdf.cell(0, 10, "Audio Transcript & Analysis Report", ln=True, align="C")
+    pdf.set_font("DejaVu", size=12)
+    pdf.ln(5)
+    pdf.cell(0, 10, f"Detected Language: {language}", ln=True)
+    pdf.ln(5)
+    pdf.set_font("DejaVu", "B", 12)
+    pdf.cell(0, 10, "English Transcript:", ln=True)
+    pdf.set_font("DejaVu", size=12)
+    very_safe_multicell(pdf, transcript_en or "", maxlen=50)
+    pdf.ln(3)
+    pdf.set_font("DejaVu", "B", 12)
+    pdf.cell(0, 10, "Brands Detected:", ln=True)
+    pdf.set_font("DejaVu", size=12)
+    very_safe_multicell(pdf, ", ".join(brands), maxlen=50)
+    pdf.set_font("DejaVu", "B", 12)
+    pdf.cell(0, 10, "Key Topics:", ln=True)
+    pdf.set_font("DejaVu", size=12)
+    very_safe_multicell(pdf, ", ".join(topics), maxlen=50)
+    pdf.set_font("DejaVu", "B", 12)
+    pdf.cell(0, 10, "Summary (Bulleted):", ln=True)
+    pdf.set_font("DejaVu", size=10)
+    for takeaway in key_takeaways.split('\n'):
+        very_safe_multicell(pdf, takeaway, maxlen=50)
+    pdf_file = "/tmp/analysis_report.pdf"
+    pdf.output(pdf_file)
+    return pdf_file
+
+def process_audio(audio_path):
+    if not audio_path or not isinstance(audio_path, str):
+        return ("No audio file provided.", "", "", "", "", "", None)
+    try:
+        with open(audio_path, "rb") as audio_file:
+            transcript = openai.audio.transcriptions.create(
+                model="whisper-1",
+                file=audio_file,
+                response_format="text"
+            )
+        transcript = make_str(transcript).strip()
+    except Exception as e:
+        return (f"Error in transcription: {e}", "", "", "", "", "", None)
+    try:
+        detected_lang = detect(transcript)
+        lang_text = {'en': 'English', 'hi': 'Hindi', 'ta': 'Tamil'}.get(detected_lang, detected_lang)
+    except Exception:
+        lang_text = "unknown"
+    transcript_en = transcript
+    if detected_lang != "en":
+        try:
+            with open(audio_path, "rb") as audio_file:
+                transcript_en = openai.audio.translations.create(
+                    model="whisper-1",
+                    file=audio_file,
+                    response_format="text"
+                )
+            transcript_en = make_str(transcript_en).strip()
+        except Exception as e:
+            transcript_en = f"Error translating: {e}"
+    try:
+        summary_obj = summarizer(transcript_en, max_length=100, min_length=30, do_sample=False)
+        summary = summary_obj[0]["summary_text"] if isinstance(summary_obj, list) and "summary_text" in summary_obj[0] else make_str(summary_obj)
+    except Exception as e:
+        summary = f"Error summarizing: {e}"
+    brands = extract_brands(transcript_en)
+    topics = extract_topics(transcript_en)
+    key_takeaways = make_bullets(summary)
+    pdf_file = create_pdf_report(lang_text, transcript_en, brands, topics, key_takeaways)
+    return (
+        lang_text,
+        transcript,
+        transcript_en,
+        ", ".join(brands),
+        ", ".join(topics),
+        key_takeaways,
+        pdf_file
+    )
+
+iface = gr.Interface(
+    fn=process_audio,
+    inputs=gr.Audio(type="filepath", label="Upload MP3/WAV Audio"),
+    outputs=[
+        gr.Textbox(label="Detected Language"),
+        gr.Textbox(label="Original Transcript"),
+        gr.Textbox(label="English Transcript (if translated)"),
+        gr.Textbox(label="Brands Detected"),
+        gr.Textbox(label="Key Topics"),
+        gr.Textbox(label="Bulleted Key Takeaways"),
+        gr.File(label="Download PDF Report")
+    ],
+    title="Audio Transcript, Brand & Topic Analysis (OpenAI Whisper + Unicode PDF Download)",
+    description="Upload your audio file (MP3/WAV). Get transcript, summary, brand & topic detection, and download PDF. Unicode (Indian language/emoji) supported."
+)
+
+iface.launch()

gitattributes

@@ -0,0 +1,36 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text
+DejaVuSans.ttf filter=lfs diff=lfs merge=lfs -text

requirements.txt

@@ -0,0 +1,7 @@
+openai
+gradio>=4.44.1
+langdetect
+transformers
+torch
+fpdf2
+keybert