Update app.py

app.py

@@ -8,7 +8,7 @@ import pandas as pd
 # ─────────────────────────────────────────
 # Config
 # ─────────────────────────────────────────
-DEFAULT_MODEL = "llama-3.3-70b-versatile"  # Groq
+DEFAULT_MODEL = "llama-3.3-70b-versatile"
 STOPWORDS = set("""
 a an and the or for nor but so yet of to in on with at by from as is are was were be being been
 i you he she it we they them us our your their this that these those here there
@@ -25,7 +25,7 @@ except ImportError:
 def get_groq_client():
     api_key = os.getenv("GROQ_API_KEY")
     if not api_key:
-        raise RuntimeError("Missing GROQ_API_KEY. Set in Space → Settings → Variables & Secrets.")
+        raise RuntimeError("Missing GROQ_API_KEY. Set it in Space → Settings → Variables & Secrets.")
     if Groq is None:
         raise RuntimeError("Package 'groq' not installed. Add 'groq' to requirements.txt.")
     return Groq(api_key=api_key)
@@ -35,7 +35,7 @@ def groq_chat(prompt, model, temperature, top_p, max_tokens):
     resp = client.chat.completions.create(
         model=model,
         messages=[
-            {"role": "system", "content": "You craft concise, original, high-signal LinkedIn posts."},
+            {"role": "system", "content": "You craft concise, original, high-signal LinkedIn posts. Respond with plain text only."},
             {"role": "user", "content": prompt}
         ],
         temperature=temperature,
@@ -45,7 +45,7 @@ def groq_chat(prompt, model, temperature, top_p, max_tokens):
     return resp.choices[0].message.content.strip()
 
 # ─────────────────────────────────────────
-# Utilities
+# Utils
 # ─────────────────────────────────────────
 def clamp(n, lo, hi):
     return max(lo, min(hi, n))
@@ -76,7 +76,7 @@ def strip_labels(text: str) -> str:
     return "\n".join(cleaned).strip()
 
 # ─────────────────────────────────────────
-# Dataset ingest + keywords (optional, improves relevance)
+# Dataset ingest + keywords (optional)
 # ─────────────────────────────────────────
 def load_posts_from_file(file) -> pd.DataFrame:
     name = file.name.lower()
@@ -88,14 +88,14 @@ def load_posts_from_file(file) -> pd.DataFrame:
         raise ValueError("Upload CSV or JSON.")
     cand = [c for c in df.columns if c.lower() in ("text","post","content","body")]
     if not cand:
-        raise ValueError("Dataset must contain a 'text' (or post/content/body) column.")
+        raise ValueError("Dataset must contain 'text' (or post/content/body).")
     if "text" not in df.columns:
         df["text"] = df[cand[0]]
     df["text"] = df["text"].fillna("").astype(str)
     return df[["text"]]
 
 def simple_rake(text, min_len=2, max_len=3, top_k=12):
-    words = re.findall(r"[A-Za-z0-9#+\-_/']+", text.lower())
+    words = re.findall(r"[A-Za-z0-9#+\\-_/']+", text.lower())
     phrases, cur = [], []
     for w in words:
         if w in STOPWORDS:
@@ -116,21 +116,21 @@ def simple_rake(text, min_len=2, max_len=3, top_k=12):
     for ph in phrases:
         s = 0.0
         for t in ph.split():
-            s += (degree.get(t,0)+1)/ (freq.get(t,1))
-        scores[ph] = scores.get(ph,0)+s
+            s += (degree.get(t,0)+1) / (freq.get(t,1))
+        scores[ph] = scores.get(ph,0) + s
     ranked = sorted(scores.items(), key=lambda x: x[1], reverse=True)
     filtered = [p for p,_ in ranked if min_len <= len(p.split()) <= max_len]
     return filtered[:top_k]
 
 def tfidf_builder(texts, top_k=8):
-    docs = [re.findall(r"[A-Za-z0-9#+\-_/']+", t.lower()) for t in texts]
+    docs = [re.findall(r"[A-Za-z0-9#+\\-_/']+", t.lower()) for t in texts]
     vocab = {}
     for d in docs:
         for w in set(d):
             vocab[w] = vocab.get(w,0)+1
     N = len(docs)
     def score(text):
-        doc = re.findall(r"[A-Za-z0-9#+\-_/']+", text.lower())
+        doc = re.findall(r"[A-Za-z0-9#+\\-_/']+", text.lower())
         tf = {}
         for w in doc:
             tf[w] = tf.get(w,0)+1
@@ -154,71 +154,94 @@ def extract_keywords(topic, df: pd.DataFrame|None):
         raw = simple_rake(topic, min_len=1, max_len=2, top_k=8)
     seen, out = set(), []
     for k in raw:
-        k2 = re.sub(r"\s+"," ",k.strip().lower())
+        k2 = re.sub(r"\\s+"," ",k.strip().lower())
         if k2 and k2 not in seen:
             seen.add(k2); out.append(k2)
     return out[:12]
 
 # ─────────────────────────────────────────
-# Stage‑2 Prompt (hidden structure, plain output)
+# Interactive clarifier
 # ─────────────────────────────────────────
-def build_stage2_prompt(topic, language, target_len, tone, keywords=None, style_cues=None):
+def need_clarification(purpose, evidence):
+    questions = []
+    if not purpose:
+        questions.append("What outcome do you want from this post? (awareness, demo requests, hiring, launch, opinion, lesson)")
+    if not evidence:
+        questions.append("Share one concrete detail to include (metric, anecdote, quote, or specific example).")
+    return questions
+
+# ─────────────────────────────────────────
+# Prompt (single post, plain text)
+# ─────────────────────────────────────────
+def build_prompt(topic, language, tone, target_len, purpose, audience, evidence, keywords, style_cues, clarifier_notes):
     kw_block = ", ".join((keywords or [])[:8]) if keywords else "N/A"
-    cues_block = "\n".join(f"- {c}" for c in (style_cues or [])[:4]) if style_cues else "- None"
+    cues_block = "\\n".join(f"- {c}" for c in (style_cues or [])[:4]) if style_cues else "- None"
+    notes = (clarifier_notes or "").strip()
     return (
-        "You are a senior LinkedIn content strategist.\n"
-        "Objective: Write a viral, insightful LinkedIn post as plain text only (no section headers, no labels).\n\n"
+        "You are a senior LinkedIn content strategist. "
+        "Write one viral, insightful LinkedIn post as plain text only (no section headers, no labels).\n\n"
         f"Language: {language}\n"
         f"Topic: \"{topic}\"\n"
-        f"Tone: \"{tone}\"\n"
+        f"Purpose: {purpose or 'awareness'}\n"
+        f"Audience: {audience or 'general professionals'}\n"
+        f"Tone: {tone}\n"
         f"Approx length: ~{target_len} words\n"
         f"Keywords to weave in naturally: {kw_block}\n"
         "Style cues (apply silently):\n"
         f"{cues_block}\n\n"
-        "Apply without mentioning rules:\n"
-        "- Curiosity‑driven first line.\n"
+        "User-provided evidence/details (incorporate if relevant):\n"
+        f"{evidence or 'None'}\n\n"
+        "Additional notes from clarifier (apply silently):\n"
+        f"{notes or 'None'}\n\n"
+        "Rules (do not mention these explicitly):\n"
+        "- Curiosity-driven first line.\n"
         "- Short paragraphs; concrete, novel insights (3–5), examples welcome.\n"
-        "- Max 2 emojis; 2–4 niche hashtags only at very end (optional).\n"
+        "- Max 2 emojis; 2–4 niche hashtags only at end (optional).\n"
         "- No repeated sentences; avoid clichés.\n"
-        "- Output must be one cohesive post in plain text. No labels or headings."
+        "- Return a single cohesive post in plain text only."
     )
 
 # ─────────────────────────────────────────
-# UI
+# Streamlit UI
 # ─────────────────────────────────────────
-st.set_page_config(page_title="LinkedIn Post Generator — Stage 2 (Groq)", layout="centered")
-st.title("Stage 2: Topic → Prompt → Llama‑3.x (Groq) → 3 Variants")
+st.set_page_config(page_title="LinkedIn Post Generator — Groq (Interactive)", layout="centered")
+st.title("LinkedIn Post Generator — Interactive (Groq)")
 
 with st.sidebar:
     st.subheader("Groq & Decoding")
-    model = st.selectbox(
-        "Groq model",
-        options=["llama-3.3-70b-versatile","llama-3.1-8b-instant","mixtral-8x7b-32768"],
-        index=0
-    )
+    model = st.selectbox("Groq model",
+        ["llama-3.3-70b-versatile","llama-3.1-8b-instant","mixtral-8x7b-32768"], index=0)
     temperature = st.slider("Temperature", 0.1, 1.2, 0.6, 0.05)
     top_p = st.slider("Top‑p", 0.1, 1.0, 0.9, 0.05)
     target_len = st.slider("Target length (words)", 60, 300, 140, 10)
     st.markdown("Set GROQ_API_KEY in Space → Settings → Variables & Secrets.")
 
-with st.form("stage2_form"):
+with st.form("main"):
     topic = st.text_input("Topic", "Generative AI for Business")
+    purpose = st.selectbox("Purpose", ["", "awareness", "lead-gen", "hiring", "product launch", "opinion", "lesson learned"], index=0)
+    audience = st.text_input("Audience", "Startup founders")
+    tone = st.selectbox("Tone", ["Professional", "Friendly", "Contrarian", "Technical", "Inspirational"], index=0)
     language = st.selectbox("Language", ["English","Urdu","Arabic","French","Spanish"], index=0)
-    tone = st.selectbox("Tone", ["Professional","Friendly","Inspirational","Technical","Concise"], index=0)
 
-    st.markdown("Optional: upload a dataset of past LinkedIn posts (CSV/JSON) with a 'text' column.")
-    uploaded = st.file_uploader("Upload CSV/JSON", type=["csv","json"])
+    st.markdown("Optional: upload CSV/JSON of past posts (must include 'text').")
+    uploaded = st.file_uploader("Upload dataset", type=["csv","json"])
 
-    st.markdown("Optional: add up to 4 style cues (one per line).")
+    st.markdown("Optional: style cues (max 4, one per line).")
     style_text = st.text_area("Style cues", value="", placeholder="Short hooks\nActionable bullets\nStories with numbers\nTactical CTA")
 
-    submitted = st.form_submit_button("Generate 3 Variants")
+    st.markdown("Optional: evidence to include (metric, anecdote, quote).")
+    evidence = st.text_area("Evidence", value="")
 
-if submitted:
-    if not os.getenv("GROQ_API_KEY"):
-        st.error("GROQ_API_KEY missing. Add it in Space → Settings → Variables & Secrets.")
-        st.stop()
+    submitted = st.form_submit_button("Continue")
+
+# Session state for clarifier & output
+if "clarifier_notes" not in st.session_state:
+    st.session_state.clarifier_notes = ""
+if "last_post" not in st.session_state:
+    st.session_state.last_post = ""
 
+if submitted:
+    # Load dataset and extract keywords
     posts_df = None
     if uploaded is not None:
         try:
@@ -226,27 +249,79 @@ if submitted:
         except Exception as e:
             st.error(f"Dataset error: {e}")
             st.stop()
-
     keywords = extract_keywords(topic, posts_df)
     style_cues = [s.strip() for s in style_text.splitlines() if s.strip()][:4]
 
-    prompt = build_stage2_prompt(topic, language, target_len, tone, keywords, style_cues)
-
-    st.subheader("Variants")
-    variants = []
-    with st.spinner("Generating with Groq..."):
-        try:
-            max_tokens = clamp(int(target_len*1.6)+120, 200, 1200)
-            # Generate 3 separate candidates
-            for i in range(3):
+    # Clarifier
+    qs = need_clarification(purpose, evidence)
+    if qs:
+        st.info("Clarifier")
+        for q in qs:
+            ans = st.text_input(q, key=f"q_{q}")
+            if ans:
+                st.session_state.clarifier_notes += f"{q} -> {ans}\n"
+        if st.button("Generate Post"):
+            prompt = build_prompt(topic, language, tone, target_len, purpose, audience, evidence, keywords, style_cues, st.session_state.clarifier_notes)
+            with st.spinner("Generating with Groq..."):
+                try:
+                    max_tokens = clamp(int(target_len*1.6)+120, 200, 1200)
+                    raw = groq_chat(prompt, model, temperature, top_p, max_tokens)
+                    clean = dedupe_sentences(strip_labels(raw))
+                    st.session_state.last_post = clean
+                except Exception as e:
+                    st.error(f"Groq generation failed: {e}")
+        # show output if available
+        if st.session_state.last_post:
+            st.subheader("Post")
+            st.write(st.session_state.last_post)
+            st.download_button("Download (.txt)", st.session_state.last_post, file_name="linkedin_post.txt")
+    else:
+        # Generate directly
+        prompt = build_prompt(topic, language, tone, target_len, purpose, audience, evidence, keywords, style_cues, st.session_state.clarifier_notes)
+        with st.spinner("Generating with Groq..."):
+            try:
+                max_tokens = clamp(int(target_len*1.6)+120, 200, 1200)
                 raw = groq_chat(prompt, model, temperature, top_p, max_tokens)
                 clean = dedupe_sentences(strip_labels(raw))
-                variants.append(clean)
+                st.session_state.last_post = clean
+            except Exception as e:
+                st.error(f"Groq generation failed: {e}")
+
+        if st.session_state.last_post:
+            st.subheader("Post")
+            st.write(st.session_state.last_post)
+            st.download_button("Download (.txt)", st.session_state.last_post, file_name="linkedin_post.txt")
+
+# Refinements (transform the last output)
+if st.session_state.last_post:
+    st.markdown("---")
+    st.subheader("Refine")
+    col1, col2, col3, col4, col5 = st.columns(5)
+    def refine(op):
+        if not st.session_state.last_post:
+            return
+        instr = {
+            "shorter": "Shorten to ~120 words. Keep the opening intact. Return plain text only.",
+            "punchier": "Make the hook more punchy and contrarian; keep total length similar. Plain text only.",
+            "add_data": "Add one concrete metric or example to support the main claim. Plain text only.",
+            "less_emoji": "Remove emojis entirely. Plain text only.",
+            "add_tags": "Append 2–4 niche hashtags at the end (new line). Plain text only."
+        }[op]
+        prompt = (
+            "You are editing a LinkedIn post. Apply the instruction and return plain text only.\n\n"
+            f"Instruction: {instr}\n\n"
+            f"Post:\n{st.session_state.last_post}"
+        )
+        try:
+            raw = groq_chat(prompt, model, temperature, top_p, clamp(600, 200, 1200))
+            st.session_state.last_post = dedupe_sentences(strip_labels(raw))
         except Exception as e:
-            st.error(f"Groq generation failed: {e}")
-            st.stop()
+            st.error(f"Refinement failed: {e}")
+
+    if col1.button("Shorter"): refine("shorter")
+    if col2.button("Punchier hook"): refine("punchier")
+    if col3.button("Add data point"): refine("add_data")
+    if col4.button("No emojis"): refine("less_emoji")
+    if col5.button("Add hashtags"): refine("add_tags")
 
-    for i, v in enumerate(variants, start=1):
-        st.markdown(f"### Post {i}")
-        st.write(v)
-        st.download_button(f"Download Post {i}", v, file_name=f"post_{i}.txt")
+    st.write(st.session_state.last_post)