Update app.py

app.py

@@ -251,83 +251,111 @@ def generate_image_from_prompt(prompt, style):
         st.warning(f"Illustrative image generation failed: {e}. Using placeholder.")
         return Image.new("RGB", (WIDTH, HEIGHT), color=(230, 230, 230))
 
+
 # ─────────────────────────────────────────────────────────────────────────────
-# REPORT GENERATION (UNCHANGED)
+# REPORT GENERATION (unchanged models – prompt now local)
 # ─────────────────────────────────────────────────────────────────────────────
-
-report_prompt = (
-    "You are a senior business analyst. Write an executive-level Markdown report "
-    "with insights & recommendations.\n"
-    "When you need a visual, insert a tag like\n"
-    '<generate_chart: "pie | sales by region">\n'
-    "(chart_type **first**, pipe, then short description). "
-    "Valid chart_type values: line, bar, scatter, pie, hist.\n"
-    f"Data Context: {json.dumps(ctx_dict, indent=2)}"
-)
-
 def generate_report_assets(key, buf, name, ctx):
     df, err = load_dataframe_safely(buf, name)
-    if err: st.error(err); return None
+    if err:
+        st.error(err)
+        return None
+
+    llm = ChatGoogleGenerativeAI(
+        model="gemini-2.0-flash", google_api_key=API_KEY, temperature=0.1
+    )
+
+    # build context dict **after** df exists
+    ctx_dict = {
+        "shape": df.shape,
+        "columns": list(df.columns),
+        "user_ctx": ctx or "General business analysis",
+    }
+
+    report_prompt = (
+        "You are a senior business analyst. Write an executive-level Markdown report "
+        "with insights & recommendations.\n"
+        'When you need a visual, insert a tag like <generate_chart: "pie | sales by region"> '
+        "(chart_type first, then a description). "
+        "Valid chart_type values: line, bar, scatter, pie, hist.\n"
+        f"Data Context: {json.dumps(ctx_dict, indent=2)}"
+    )
 
-    llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash",
-                                 google_api_key=API_KEY, temperature=0.1)
-    ctx_dict = {"shape": df.shape, "columns": list(df.columns),
-                "user_ctx": ctx or "General business analysis"}
     md = llm.invoke(report_prompt).content
 
-    # Replace tags with static charts
+    # ------------------------------------------------------------------ charts
     chart_descs = extract_chart_tags(md)[:MAX_CHARTS]
     charts = {}
     if chart_descs:
-        ag = create_pandas_dataframe_agent(llm=llm, df=df, verbose=False, allow_dangerous_code=True)
+        agent = create_pandas_dataframe_agent(
+            llm=llm, df=df, verbose=False, allow_dangerous_code=True
+        )
         for d in chart_descs:
             with st.spinner(f"Generating chart: {d}"):
                 with plt.ioff():
                     try:
-                        ag.run(f"Create a {d} with Matplotlib and save."); fig = plt.gcf()
+                        agent.run(f"Create a {d} with Matplotlib and save.")
+                        fig = plt.gcf()
                         if fig.axes:
                             p = Path(tempfile.gettempdir()) / f"{uuid.uuid4()}.png"
                             fig.savefig(p, dpi=300, bbox_inches="tight", facecolor="white")
                             charts[d] = str(p)
                         plt.close("all")
-                    except: plt.close("all")
+                    except Exception:
+                        plt.close("all")
 
     preview = repl_tags(
-        md, charts,
-        lambda p: f'<img src="data:image/png;base64,{base64.b64encode(Path(p).read_bytes()).decode()}" '
-                  f'style="max-width:100%;">'
+        md,
+        charts,
+        lambda p: f'<img src="data:image/png;base64,'
+        f'{base64.b64encode(Path(p).read_bytes()).decode()}">'
     )
     pdf = build_pdf(md, charts)
-    return {"type": "report", "preview": preview, "pdf": pdf, "report_md": md, "key": key}
+
+    return {
+        "type": "report",
+        "preview": preview,
+        "pdf": pdf,
+        "report_md": md,
+        "key": key,
+    }
+
 
 # ─────────────────────────────────────────────────────────────────────────────
-# VIDEO GENERATION (ANIMATED CHARTS)
+# VIDEO GENERATION (animated charts – prompt now local)
 # ─────────────────────────────────────────────────────────────────────────────
-# ── Video script prompt ───────────────────────────────────
-story_prompt = (
-    f"Create a script for a short business video with exactly {VIDEO_SCENES} scenes.\n"
-    "For each scene:\n"
-    "1. Provide 1–2 sentences of narration.\n"
-    '2. If a visual is useful, add <generate_chart: "bar | monthly revenue"> '
-    "(chart_type first).\n"
-    "3. Separate scenes with [SCENE_BREAK].\n"
-    f"Data Context: {json.dumps(ctx_dict, indent=2)}"
-)
-
 def generate_video_assets(key, buf, name, ctx, style, animate_charts=True):
-    # FFmpeg presence
     try:
         subprocess.run(["ffmpeg", "-version"], check=True, capture_output=True)
     except Exception:
-        st.error("🔴 FFmpeg not available — cannot render video."); return None
+        st.error("🔴 FFmpeg not available — cannot render video.")
+        return None
 
     df, err = load_dataframe_safely(buf, name)
-    if err: st.error(err); return None
+    if err:
+        st.error(err)
+        return None
+
+    llm = ChatGoogleGenerativeAI(
+        model="gemini-2.0-flash", google_api_key=API_KEY, temperature=0.2
+    )
+
+    ctx_dict = {
+        "shape": df.shape,
+        "columns": list(df.columns),
+        "user_ctx": ctx or "General business analysis",
+    }
+
+    story_prompt = (
+        f"Create a script for a short business video with exactly {VIDEO_SCENES} scenes.\n"
+        "For each scene:\n"
+        "1. Provide 1–2 sentences of narration.\n"
+        '2. If a visual is helpful, add <generate_chart: "bar | monthly revenue"> '
+        "(chart_type first).\n"
+        "3. Separate scenes with [SCENE_BREAK].\n"
+        f"Data Context: {json.dumps(ctx_dict, indent=2)}"
+    )
 
-    llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash",
-                                 google_api_key=API_KEY, temperature=0.2)
-    ctx_dict = {"shape": df.shape, "columns": list(df.columns),
-                "user_ctx": ctx or "General business analysis"}
     script = llm.invoke(story_prompt).content
     scenes = [s.strip() for s in script.split("[SCENE_BREAK]") if s.strip()]
 
@@ -335,45 +363,66 @@ def generate_video_assets(key, buf, name, ctx, style, animate_charts=True):
 
     for idx, scene in enumerate(scenes[:VIDEO_SCENES]):
         st.progress((idx + 1) / VIDEO_SCENES, text=f"Processing Scene {idx+1}/{VIDEO_SCENES}…")
+
         chart_tags = extract_chart_tags(scene)
-        narrative  = repl_tags(scene, {}, lambda _: "").strip()
+        narrative = repl_tags(scene, {}, lambda _: "").strip()
 
-        # Audio
+        # ---------------- audio -------------------------------------------
         audio_bytes, _ = deepgram_tts(narrative)
         audio_path = Path(tempfile.gettempdir()) / f"{uuid.uuid4()}.mp3"
-        (audio_path.write_bytes(audio_bytes) if audio_bytes else None)
+        if audio_bytes:
+            audio_path.write_bytes(audio_bytes)
         duration = get_audio_duration(str(audio_path)) if audio_bytes else 5.0
-        audio_parts.append(str(audio_path)); temps.append(audio_path)
+        audio_parts.append(str(audio_path))
+        temps.append(audio_path)
 
-        # Video
+        # ---------------- visual ------------------------------------------
         if chart_tags and animate_charts:
             clip_path = Path(tempfile.gettempdir()) / f"{uuid.uuid4()}.mp4"
             animate_chart(chart_tags[0], df, duration, clip_path, FPS)
-            video_parts.append(str(clip_path)); temps.append(clip_path)
+            video_parts.append(str(clip_path))
+            temps.append(clip_path)
         else:
-            # illustrative image fade
             img = generate_image_from_prompt(narrative, style)
             png_tmp = Path(tempfile.gettempdir()) / f"{uuid.uuid4()}.png"
-            img.save(png_tmp); temps.append(png_tmp)
+            img.save(png_tmp)
+            temps.append(png_tmp)
             clip_path = Path(tempfile.gettempdir()) / f"{uuid.uuid4()}.mp4"
             animate_image_fade(
                 cv2.cvtColor(np.array(img.resize((WIDTH, HEIGHT))), cv2.COLOR_RGB2BGR),
-                duration, clip_path, FPS
+                duration,
+                clip_path,
+                FPS,
             )
-            video_parts.append(str(clip_path)); temps.append(clip_path)
+            video_parts.append(str(clip_path))
+            temps.append(clip_path)
 
-    # Concatenate media
+    # -------------- concatenate ----------------------------------------------
     silent_vid = Path(tempfile.gettempdir()) / f"{uuid.uuid4()}.mp4"
     concat_media(video_parts, silent_vid, "video")
-    audio_mix  = Path(tempfile.gettempdir()) / f"{uuid.uuid4()}.mp3"
+    audio_mix = Path(tempfile.gettempdir()) / f"{uuid.uuid4()}.mp3"
     concat_media(audio_parts, audio_mix, "audio")
 
-    final_vid  = Path(tempfile.gettempdir()) / f"{key}.mp4"
+    final_vid = Path(tempfile.gettempdir()) / f"{key}.mp4"
     subprocess.run(
-        ["ffmpeg", "-y", "-i", str(silent_vid), "-i", str(audio_mix),
-         "-c:v", "copy", "-c:a", "aac", "-shortest", str(final_vid)],
-        check=True, capture_output=True
+        [
+            "ffmpeg",
+            "-y",
+            "-i",
+            str(silent_vid),
+            "-i",
+            str(audio_mix),
+            "-c:v",
+            "copy",
+            "-c:a",
+            "aac",
+            "-shortest",
+            str(final_vid),
+        ],
+        check=True,
+        capture_output=True,
     )
+
     return {"type": "video", "video_path": str(final_vid), "key": key}
 
 # ─────────────────────────────────────────────────────────────────────────────