Update app.py

app.py

@@ -1,260 +1,186 @@
 ###############################################################################
-# app.py  –  streamlined & stable
+# app.py — stable, cached, with on-demand TTS
 ###############################################################################
-import os, re, json, uuid, tempfile, asyncio, base64, io, hashlib
+import os, re, json, uuid, asyncio, base64, io, hashlib, tempfile
 from pathlib import Path
 import streamlit as st; st.set_page_config("AI Report & Slides", layout="wide")
-import pandas as pd, matplotlib; matplotlib.use("Agg")
-import matplotlib.pyplot as plt
+import pandas as pd, matplotlib; matplotlib.use("Agg"); import matplotlib.pyplot as plt
 from fpdf import FPDF, HTMLMixin
 from markdown import markdown
 try: from markdown_it import MarkdownIt
 except ImportError: MarkdownIt = None
 from pptx import Presentation
 from pptx.util import Inches, Pt
-
-# Google APIs
-from google.oauth2.service_account import Credentials
-from googleapiclient.discovery import build
-
-# Gemini / ADK
 from google import genai
 from google.genai import types
-from google.adk.agents   import LlmAgent, SequentialAgent
-from google.adk.runners  import Runner
+from google.adk.agents import LlmAgent, SequentialAgent
+from google.adk.runners import Runner
 from google.adk.sessions import InMemorySessionService
 from langchain_experimental.agents import create_pandas_dataframe_agent
-from langchain_google_genai       import ChatGoogleGenerativeAI
+from langchain_google_genai import ChatGoogleGenerativeAI
 
-# ─────────────────────────────────────────────────────────────────────────────
-# CONFIG
-# ─────────────────────────────────────────────────────────────────────────────
+# ————————————————————————————————————————————————————————————— CONFIG
 FONT_DIR = Path(__file__).parent
-FONT_REG = FONT_DIR / "NotoSans-Regular.ttf"
-FONT_BLD = FONT_DIR / "NotoSans-Bold.ttf"
-FONT_NAME = "NotoSans"
-SLIDE_COUNT = 7
-TTS_MODEL  = "gemini-2.5-flash-preview-tts"
-CREDS_JSON = os.getenv("GOOGLE_SERVICE_ACCOUNT_JSON")   # optional
-
-os.environ["STREAMLIT_CONFIG_DIR"] = tempfile.gettempdir()
-os.environ["MPLCONFIGDIR"]         = tempfile.gettempdir()
+FONT_REG = FONT_DIR/"NotoSans-Regular.ttf"; FONT_BLD = FONT_DIR/"NotoSans-Bold.ttf"
+FONT = "NotoSans"; SLIDES = 7
+TTS_MODEL = "gemini-2.5-flash-preview-tts"
 
-API_KEY = os.getenv("GEMINI_API_KEY")
+API_KEY = os.getenv("GEMINI_API_KEY"); st.experimental_set_query_params()
 if not API_KEY: st.error("Set GEMINI_API_KEY"); st.stop()
-GEMINI = genai.Client(api_key=API_KEY)
-
-# ─────────────────────────────────────────────────────────────────────────────
-# SMALL HELPERS
-# ─────────────────────────────────────────────────────────────────────────────
-fix_bullets = lambda txt: txt.replace("\x95", "•")
-
-def hash_df(df: pd.DataFrame) -> str:
-    return hashlib.sha1(pd.util.hash_pandas_object(df, index=True).values).hexdigest()
-
-def pcm_to_wav(raw: bytes, rate=24_000):
-    import wave, contextlib, io as _io
-    buf=_io.BytesIO()
-    with contextlib.closing(wave.open(buf, "wb")) as f:
-        f.setnchannels(1); f.setsampwidth(2); f.setframerate(rate); f.writeframes(raw)
+GEM = genai.Client(api_key=API_KEY)
+
+# ————————————————————————————————————————————————————————————— HELPERS
+fix_bullet = lambda t: t.replace("\x95","•")
+def pcm_to_wav(raw, rate=24_000):
+    import wave, contextlib; buf=io.BytesIO()
+    with contextlib.closing(wave.open(buf,'wb')) as w:
+        w.setnchannels(1); w.setsampwidth(2); w.setframerate(rate); w.writeframes(raw)
     return buf.getvalue()
-
-# ─────────────────────────────────────────────────────────────────────────────
-#   CACHED BUILDERS
-# ─────────────────────────────────────────────────────────────────────────────
+def df_hash(df): return hashlib.sha1(pd.util.hash_pandas_object(df, index=True).values).hexdigest()
+
+@st.cache_data(show_spinner=False)
+def tts_cached(text):
+    txt = text[:500]+"…" if len(text)>500 else text
+    try:
+        r=GEM.models.generate_content(
+            model=TTS_MODEL, contents=f"Say clearly: {txt}",
+            config=types.GenerateContentConfig(
+                response_modalities=["AUDIO"],
+                speech_config=types.SpeechConfig(
+                    voice_config=types.VoiceConfig(
+                        prebuilt_voice_config=types.PrebuiltVoiceConfig(voice_name="Kore")))))
+        audio=r.candidates[0].content.parts[0].inline_data
+        return audio.data, audio.mime_type
+    except Exception as e: st.error(f"TTS error: {e}"); return None,None
+
+# — cached heavy builders —
 @st.cache_resource(show_spinner=False)
-def pandas_agent(df_hash: str):
-    return create_pandas_dataframe_agent(
-        llm=ChatGoogleGenerativeAI(model="gemini-2.0-flash", google_api_key=API_KEY),
-        df=st.session_state.df,
-        verbose=False, allow_dangerous_code=True)
+def pandas_agent(dfh): return create_pandas_dataframe_agent(
+        llm=ChatGoogleGenerativeAI(model="gemini-2.0-flash",google_api_key=API_KEY),
+        df=st.session_state.df, verbose=False, allow_dangerous_code=True)
 
 @st.cache_resource(show_spinner=False)
-def build_charts(descs: tuple, df_hash: str):
-    agent = pandas_agent(df_hash)
-    out={}
+def charts(descs, dfh):
+    pa=pandas_agent(dfh); mp={}
     for d in descs:
-        if d in out: continue
-        agent.run(f"Create a {d} using matplotlib"); fig=plt.gcf()
+        pa.run(f"Create a {d} using matplotlib")
         p=Path(tempfile.gettempdir())/f"{uuid.uuid4()}.png"
-        fig.savefig(p,dpi=300,bbox_inches="tight"); plt.close()
-        out[d]=str(p)
-    return out
+        plt.gcf().savefig(p,dpi=300,bbox_inches="tight"); plt.close()
+        mp[d]=str(p)
+    return mp
 
 @st.cache_resource(show_spinner=False)
-def build_pdf(html_ready_md: str) -> bytes:
-    html_ready_md = fix_bullets(html_ready_md)
-    html = (MarkdownIt("commonmark",{"breaks":True}).enable("table").render(html_ready_md)
-            if MarkdownIt else markdown(html_ready_md,extensions=["tables"]))
-    dummy=FPDF(); page_w=dummy.w-2*dummy.l_margin; px=lambda mm:int(mm*dummy.k)
+def build_pdf(html_md):
+    html_md=fix_bullet(html_md)
+    html=(MarkdownIt("commonmark",{"breaks":True}).enable("table").render(html_md)
+          if MarkdownIt else markdown(html_md,extensions=["tables"]))
+    dummy=FPDF(); width=dummy.w-2*dummy.l_margin; px=lambda mm:int(mm*dummy.k)
     html=re.sub(r'(<table[^>]*>)',r'\1<font size="8">',html).replace('</table>','</font></table>')
-    html=re.sub(r'<table(?![^>]*width=)',lambda m:f'<table width="{px(page_w)}"',html)
+    html=re.sub(r'<table(?![^>]*width=)',lambda m:f'<table width="{px(width)}"',html)
     html=re.sub(r'<img[^>]+src="([^"]+\.png)"[^>]*>',
-                lambda m:f'<img src="{m.group(1)}" width="{int(page_w)}" />',html)
+                lambda m:f'<img src="{m.group(1)}" width="{int(width)}" />',html)
     class PDF(FPDF,HTMLMixin): pass
     pdf=PDF(); pdf.set_auto_page_break(True,15)
-    for s,p in [("",FONT_REG),("B",FONT_BLD),("I",FONT_REG),("BI",FONT_BLD)]:
-        pdf.add_font(FONT_NAME,s,str(p),uni=True)
-    pdf.set_fallback_fonts([FONT_NAME]); pdf.add_page()
-    pdf.set_font(FONT_NAME,"B",18); pdf.cell(0,12,"AI-Generated Business Report",ln=True); pdf.ln(3)
-    pdf.set_font(FONT_NAME,"",11); pdf.write_html(html)
+    for s,f in [("",FONT_REG),("B",FONT_BLD),("I",FONT_REG),("BI",FONT_BLD)]: pdf.add_font(FONT,s,str(f),uni=True)
+    pdf.set_fallback_fonts([FONT]); pdf.add_page()
+    pdf.set_font(FONT,"B",18); pdf.cell(0,12,"AI-Generated Business Report",ln=True); pdf.ln(3)
+    pdf.set_font(FONT,"",11); pdf.write_html(html)
     return bytes(pdf.output(dest="S"))
 
 @st.cache_resource(show_spinner=False)
-def build_pptx(slide_texts: tuple, desc2png: dict) -> bytes:
-    prs=Presentation(); layout=prs.slide_layouts[1]      # title+content
-    for idx,raw in enumerate(slide_texts):
-        raw=fix_bullets(raw); plain=re.sub(r'<.*?>','',raw).strip()
-        title_line, *body_lines = plain.splitlines()
-        body_lines=[l.strip().lstrip('*').strip() for l in body_lines if l.strip()]
-        slide=prs.slides.add_slide(layout)
-        slide.shapes.title.text=title_line
-        tf=slide.shapes.placeholders[1].text_frame
-        tf.clear()
-        for l in body_lines:
-            p=tf.add_paragraph(); p.text=l; p.font.size=Pt(18); p.level=0
-        m=re.search(r'<generate_chart:\s*"([^"]+)"\s*>', raw)
-        if m and m.group(1) in desc2png:
-            slide.shapes.add_picture(desc2png[m.group(1)], Inches(1), Inches(4), width=Inches(8))
-    fout=io.BytesIO(); prs.save(fout); return fout.getvalue()
-
-@st.cache_resource(show_spinner=False)
-def upload_pptx_to_slides(pptx_bytes: bytes) -> str|None:
-    if not CREDS_JSON: return None
-    creds=Credentials.from_service_account_file(
-        CREDS_JSON,
-        scopes=["https://www.googleapis.com/auth/drive.file",
-                "https://www.googleapis.com/auth/presentations"])
-    drv=build("drive","v3",credentials=creds)
-    meta={"name":f"AI-Slides‐{uuid.uuid4().hex[:6]}",
-          "mimeType":"application/vnd.google-apps.presentation"}
-    media={"body":io.BytesIO(pptx_bytes),
-           "mimeType":"application/vnd.openxmlformats-officedocument.presentationml.presentation"}
-    f=drv.files().create(body=meta,media_body=media).execute()
-    return f"https://docs.google.com/presentation/d/{f['id']}"
-
-# ─────────────────────────────────────────────────────────────────────────────
-#   UI : INPUTS
-# ─────────────────────────────────────────────────────────────────────────────
-mode = st.radio("Output:", ["Report","Presentation","Both"],horizontal=True)
-file  = st.file_uploader("CSV / XLSX",["csv","xlsx"])
-context = st.text_area("Optional business context / goals")
+def build_pptx(slides, mp):
+    prs=Presentation(); layout=prs.slide_layouts[1]
+    for i,raw in enumerate(slides):
+        raw=fix_bullet(raw)
+        title,*body=filter(None,[l.strip().lstrip('*').strip() for l in re.sub(r'<.*?>','',raw).splitlines()])
+        sl=prs.slides.add_slide(layout); sl.shapes.title.text=title
+        tf=sl.placeholders[1].text_frame; tf.clear()
+        for line in body: p=tf.add_paragraph(); p.text=line; p.font.size=Pt(24)
+        m=re.search(r'<generate_chart:\s*"([^"]+)"',raw)
+        if m and m.group(1) in mp:
+            sl.shapes.add_picture(mp[m.group(1)],Inches(1),Inches(4),width=Inches(8))
+    buf=io.BytesIO(); prs.save(buf); return buf.getvalue()
+
+# ————————————————————————————————————————————————————————————— UI INPUT
+mode = st.radio("Output:",["Report","Presentation","Both"],horizontal=True)
+upl  = st.file_uploader("CSV / XLSX",["csv","xlsx"])
+ctx  = st.text_area("Optional business context")
 if not st.button("🚀 Generate"): st.stop()
-if not file: st.warning("Upload data first"); st.stop()
-
-# ─────────────────────────────────────────────────────────────────────────────
-#   LOAD DATAFRAME
-# ─────────────────────────────────────────────────────────────────────────────
-try:
-    st.session_state.df = pd.read_excel(file) if file.name.lower().endswith(".xlsx") else pd.read_csv(file)
-except Exception as e:
-    st.error(f"Failed to read file: {e}"); st.stop()
-df_hash = hash_df(st.session_state.df)
-st.success(f"Loaded {st.session_state.df.shape[0]} rows × {st.session_state.df.shape[1]} cols")
-
-# ─────────────────────────────────────────────────────────────────────────────
-#   BUILD & RUN AGENTS
-# ─────────────────────────────────────────────────────────────────────────────
-subs=[]
-if mode in ("Report","Both"):
-    subs.append(LlmAgent(name="ReportAgent",model="gemini-2.5-flash",
-        instruction="Write an exec-level Markdown report, using <generate_chart:\"…\"> placeholders."))
-if mode in ("Presentation","Both"):
-    subs.append(LlmAgent(name="PresentationAgent",model="gemini-2.5-flash",
-        instruction=f"Create {SLIDE_COUNT} slides. Each begins 'Slide X:' and may include <generate_chart:\"…\"> placeholders."))
-
-root=SequentialAgent(name="Pipeline",sub_agents=subs)
-svc=InMemorySessionService(); session_id=str(uuid.uuid4())
-asyncio.run(svc.create_session(app_name="app",user_id="usr",session_id=session_id))
-runner=Runner(agent=root,app_name="app",session_service=svc)
-
-ctx = {"dataset_info":{"shape": st.session_state.df.shape,
-                       "columns": list(st.session_state.df.columns),
-                       "dtypes": st.session_state.df.dtypes.astype(str).to_dict(),
-                       "missing": st.session_state.df.isna().sum().to_dict()},
-       "user_context": context,
-       "preview": st.session_state.df.head().to_dict()}
-
-final={}
-async def _run():
-    async for ev in runner.run_async(user_id="usr",session_id=session_id,
-                                     new_message=types.Content(role="user",
-                                                               parts=[types.Part(text=json.dumps(ctx))])):
-        if ev.is_final_response(): final[ev.author]=ev.content.parts[0].text
-asyncio.run(_run())
-
-# ─────────���───────────────────────────────────────────────────────────────────
-#   CHARTS (one build for everything)
-# ─────────────────────────────────────────────────────────────────────────────
-all_md = "\n".join(final.values())
-chart_descs = re.findall(r'<generate_chart:\s*"([^"]+)"\s*>', all_md)
-png_map = build_charts(tuple(chart_descs), df_hash)
+if not upl: st.warning("Upload data"); st.stop()
+
+# ————————————————————————————————————————————————————————————— HEAVY GEN
+if "generated" not in st.session_state:
+    try:
+        st.session_state.df = pd.read_excel(upl) if upl.name.lower().endswith(".xlsx") else pd.read_csv(upl)
+    except Exception as e: st.error(e); st.stop()
+    h = df_hash(st.session_state.df)
+    # Agents
+    agents=[]
+    if mode in ("Report","Both"):
+        agents.append(LlmAgent(name="ReportAgent",model="gemini-2.5-flash",
+                               instruction="Write exec-level Markdown report with <generate_chart:\"…\">"))
+    if mode in ("Presentation","Both"):
+        agents.append(LlmAgent(name="PresentationAgent",model="gemini-2.5-flash",
+                               instruction=f"Create {SLIDES} slides starting 'Slide X:' with <generate_chart:\"…\"> tags"))
+    root=SequentialAgent(name="root",sub_agents=agents)
+    svc=InMemorySessionService(); sid=str(uuid.uuid4())
+    asyncio.run(svc.create_session(app_name="app",user_id="u",session_id=sid))
+    runner=Runner(agent=root,app_name="app",session_service=svc)
+    context={"dataset_info":{"shape":st.session_state.df.shape},
+             "user_context":ctx,"preview":st.session_state.df.head().to_dict()}
+    result={}
+    async def _go():
+        async for ev in runner.run_async(user_id="u",session_id=sid,
+                                         new_message=types.Content(role="user",
+                                                                   parts=[types.Part(text=json.dumps(context))])):
+            if ev.is_final_response(): result[ev.author]=ev.content.parts[0].text
+    asyncio.run(_go())
+    st.session_state.result=result
+    # charts
+    tags=re.findall(r'<generate_chart:\s*"([^"]+)"',"\n".join(result.values()))
+    mp=charts(tuple(tags),h); st.session_state.chart_map=mp
+    # slides & docs
+    if "ReportAgent" in result:
+        md=result["ReportAgent"]
+        md_full=md
+        for d,p in mp.items(): md_full=re.sub(rf'<generate_chart:\s*"{re.escape(d)}"',f'<img src="{p}"',md_full)
+        st.session_state.pdf=build_pdf(md_full)
+    if "PresentationAgent" in result:
+        slides=[s.strip() for s in re.split(r'\bSlide\s+\d+:',result["PresentationAgent"]) if s.strip()][:SLIDES]
+        st.session_state.slides=slides
+        st.session_state.pptx=build_pptx(tuple(slides),mp)
+    st.session_state.idx=0
+    st.session_state.generated=True
+
+# ————————————————————————————————————————————————————————————— TABS UI
+if mode=="Both": tab_rep,tab_pre=st.tabs(["Report","Slides"])
+else: tab_rep=tab_pre=st
 
-# ─────────────────────────────────────────────────────────────────────────────
-#   TABS
-# ─────────────────────────────────────────────────────────────────────────────
-if mode=="Both":
-    t_report,t_pres=st.tabs(["📄 Report","📑 Slides"])
-else:
-    t_report=t_pres=st
-
-# ─────────────────────────────────────────────────────────────────────────────
-#   REPORT
-# ─────────────────────────────────────────────────────────────────────────────
 if mode in ("Report","Both"):
-    md_raw = fix_bullets(final["ReportAgent"])
-    md_prev=md_raw; md_pdf=md_raw
-    for d,p in png_map.items():
+    md_prev=st.session_state.result["ReportAgent"]
+    for d,p in st.session_state.chart_map.items():
         b64=base64.b64encode(open(p,"rb").read()).decode()
-        md_prev=re.sub(rf'<generate_chart:\s*"{re.escape(d)}"\s*>',
-                       f'<img src="data:image/png;base64,{b64}" style="max-width:100%;margin:6px 0;" />',
-                       md_prev)
-        md_pdf =re.sub(rf'<generate_chart:\s*"{re.escape(d)}"\s*>',
-                       f'<img src="{p}" />',md_pdf)
-    with t_report:
+        md_prev=re.sub(rf'<generate_chart:\s*"{re.escape(d)}"',f'<img src="data:image/png;base64,{b64}"',md_prev)
+    with tab_rep:
         st.markdown(md_prev,unsafe_allow_html=True)
-        if "pdf_bytes" not in st.session_state:
-            st.session_state.pdf_bytes=build_pdf(md_pdf)
-        st.download_button("⬇️ Download PDF", st.session_state.pdf_bytes,
-                           "business_report.pdf","application/pdf")
+        st.download_button("Download PDF",st.session_state.pdf,"report.pdf","application/pdf")
 
-# ─────────────────────────────────────────────────────────────────────────────
-#   PRESENTATION
-# ─────────────────────────────────────────────────────────────────────────────
 if mode in ("Presentation","Both"):
-    raw = fix_bullets(final["PresentationAgent"])
-    slides=[s.strip() for s in re.split(r'\bSlide\s+\d+:',raw) if s.strip()][:SLIDE_COUNT]
-    st.session_state.slides=slides
-    if "pptx_bytes" not in st.session_state:
-        st.session_state.pptx_bytes=build_pptx(tuple(slides), png_map)
-
-    if "idx" not in st.session_state: st.session_state.idx=0
-
-    def show_slide(i):
-        txt=slides[i]; st.markdown(f"### Slide {i+1}/{len(slides)}"); st.write(txt)
-        if st.button("🔊 Speak",key=f"tts{i}"):
-            with st.spinner("Generating TTS…"):
-                aud,mime=generate_tts_audio(GEMINI,txt)
+    with tab_pre:
+        slides=st.session_state.slides
+        idx=st.session_state.idx
+        st.markdown(f"### Slide {idx+1}/{len(slides)}")
+        st.write(fix_bullet(slides[idx]))
+        if st.button("🔊 Speak",key="speak"):
+            with st.spinner("TTS…"):
+                aud,mime=tts_cached(slides[idx])
                 if aud:
-                    st.audio(pcm_to_wav(aud) if 'pcm' in mime or 'L16' in mime else aud,
-                             format="audio/wav" if 'pcm' in mime or 'L16' in mime else mime)
-
-    with t_pres:
-        c1,c2=st.columns(2)
-        c1.button("⬅️",disabled=st.session_state.idx==0,
-                  on_click=lambda: st.session_state.__setitem__("idx",st.session_state.idx-1))
-        c2.button("➡️",disabled=st.session_state.idx==len(slides)-1,
-                  on_click=lambda: st.session_state.__setitem__("idx",st.session_state.idx+1))
-        show_slide(st.session_state.idx)
-
-        st.download_button("⬇️ PPTX", st.session_state.pptx_bytes,
+                    st.audio(pcm_to_wav(aud) if mime.startswith("audio/raw") else aud,
+                             format="audio/wav" if mime.startswith("audio/raw") else mime)
+        col1,col2=st.columns(2)
+        if col1.button("⬅️",disabled=idx==0): st.session_state.idx-=1; st.rerun()
+        if col2.button("➡️",disabled=idx==len(slides)-1): st.session_state.idx+=1; st.rerun()
+        st.download_button("Download PPTX",st.session_state.pptx,
                            "presentation.pptx",
-                           ("application/vnd.openxmlformats-officedocument."
-                            "presentationml.presentation"))
-        if st.button("⬆️ Upload to Google Slides"):
-            with st.spinner("Uploading…"):
-                url=upload_pptx_to_slides(st.session_state.pptx_bytes)
-            url and st.success(f"[Open Slides]({url})")
-###############################################################################
-# End
+                           "application/vnd.openxmlformats-officedocument.presentationml.presentation")
 ###############################################################################