Update app.py

app.py

@@ -1,7 +1,6 @@
-# app.py
 """
 Educational Pre-Audit (EPA) Generator
-Generates customized school audit reports (English Word .docx or Urdu Markdown .md)
+Generates customized English school audit reports (.docx)
 Developed by Najaf Ali Sharqi
 """
 
@@ -19,46 +18,41 @@ from docx.enum.text import WD_PARAGRAPH_ALIGNMENT, WD_LINE_SPACING
 from docx.oxml import OxmlElement
 from docx.oxml.ns import qn
 
-# Gemini configuration (optional)
-GENAI_AVAILABLE = False
-try:
-    import google.generativeai as genai
-    key = os.getenv("GOOGLE_API_KEY")
-    if key:
-        genai.configure(api_key=key)
-        GENAI_AVAILABLE = True
-except Exception:
-    GENAI_AVAILABLE = False
+# Optional: Gemini or Groq API (comment out Gemini if unused)
+from groq import Groq
+client = Groq(api_key="YOUR_GROQ_API_KEY")  # 🔑 Replace with your key
 
 
 def sanitize_text(text: str) -> str:
-    """Remove symbols and emojis"""
+    """Remove symbols and emojis."""
     return "".join(ch for ch in text if ord(ch) >= 32)
 
 
-def build_prompt(school_data: dict, language: str) -> str:
-    prompt = (
+def build_prompt(school_data: dict) -> str:
+    return (
         "You are EduAudit AI, a professional educational auditor.\n"
-        "Generate a detailed Educational Pre-Audit Report with clearly separated sections:\n"
+        "Generate a detailed Educational Pre-Audit Report with the following structure:\n"
         "1. Executive Summary\n2. Domain-wise Findings\n3. Key Findings Table\n"
         "4. Recommendations and SMART Action Plan\n5. Overall Audit Rating (1–5)\n6. Conclusion\n\n"
-        "Each heading must be bold and start on a new line. Avoid any emojis or special characters.\n"
-        "Use clear, professional tone.\n\n"
-        f"DATA:\n{json.dumps(school_data, ensure_ascii=False, indent=2)}\n\n"
-        f"Language: {language}\n"
+        "Each heading must be bold and start on a new line. Sub-headings may also be bold.\n"
+        "The normal text under headings should not be bold.\n"
+        "Use clear, professional tone and align with Pakistani educational context.\n\n"
+        f"DATA:\n{json.dumps(school_data, ensure_ascii=False, indent=2)}\n"
     )
-    return prompt
 
 
-def generate_with_gemini(prompt: str):
-    if not GENAI_AVAILABLE:
-        raise RuntimeError("Gemini API not configured.")
-    model = genai.GenerativeModel("gemini-2.5-flash")
-    r = model.generate_content(prompt)
-    return r.text
+def generate_with_groq(prompt: str) -> str:
+    """Generate text using Groq API."""
+    r = client.chat.completions.create(
+        model="llama3-8b-8192",
+        messages=[{"role": "user", "content": prompt}],
+        temperature=0.6,
+    )
+    return r.choices[0].message.content.strip()
 
 
 def set_paragraph_format(paragraph, font_size=12, spacing=1.5, bold=False):
+    """Applies uniform paragraph formatting."""
     for run in paragraph.runs:
         run.font.name = "Times New Roman"
         run.font.size = Pt(font_size)
@@ -67,8 +61,10 @@ def set_paragraph_format(paragraph, font_size=12, spacing=1.5, bold=False):
 
 
 def create_docx(report: str, data: dict, charts: list):
+    """Creates a .docx file with bold headings only."""
     doc = Document()
 
+    # Header
     section = doc.sections[0]
     header = section.header.paragraphs[0]
     header.text = "Developed by Najaf Ali Sharqi"
@@ -80,12 +76,15 @@ def create_docx(report: str, data: dict, charts: list):
     set_paragraph_format(title, bold=True)
 
     # Metadata
-    p1 = doc.add_paragraph(f"Generated on: {datetime.now():%Y-%m-%d %H:%M}")
-    p2 = doc.add_paragraph(f"Institution: {data.get('name','')}")
-    p3 = doc.add_paragraph(f"Principal: {data.get('principal','')}")
-    p4 = doc.add_paragraph(f"Location: {data.get('location','')}")
-    p5 = doc.add_paragraph(f"Level: {data.get('level','')}")
-    for p in [p1, p2, p3, p4, p5]:
+    metadata = [
+        f"Generated on: {datetime.now():%Y-%m-%d %H:%M}",
+        f"Institution: {data.get('name','')}",
+        f"Principal: {data.get('principal','')}",
+        f"Location: {data.get('location','')}",
+        f"Level: {data.get('level','')}",
+    ]
+    for line in metadata:
+        p = doc.add_paragraph(line)
         set_paragraph_format(p)
 
     doc.add_paragraph()
@@ -97,26 +96,33 @@ def create_docx(report: str, data: dict, charts: list):
         cap.alignment = WD_PARAGRAPH_ALIGNMENT.CENTER
         set_paragraph_format(cap)
 
-    # Split report by headings
+    # --- Text formatting logic ---
     lines = report.split("\n")
     for line in lines:
-        if not line.strip():
+        line = line.strip()
+        if not line:
             continue
-        if line.strip().endswith(":") or line.strip().endswith("۔") or line.strip().endswith("."):
-            para = doc.add_paragraph(line.strip())
-            para.style = "Normal"
+
+        # Detect headings / subheadings (only bold)
+        if (
+            line.lower().startswith(("executive summary", "domain-wise", "key findings",
+                                     "recommendations", "overall audit", "conclusion"))
+            or line.strip().endswith(":")
+        ):
+            para = doc.add_paragraph(line)
             set_paragraph_format(para, bold=True)
+            doc.add_paragraph()  # blank line after heading
         else:
-            para = doc.add_paragraph(line.strip())
-            para.style = "Normal"
-            set_paragraph_format(para)
+            para = doc.add_paragraph(line)
+            set_paragraph_format(para, bold=False)
 
-    # Footer with page number
+    # Footer with page numbers
     footer = section.footer.paragraphs[0]
-    fld = OxmlElement('w:fldSimple')
-    fld.set(qn('w:instr'), 'PAGE')
+    fld = OxmlElement("w:fldSimple")
+    fld.set(qn("w:instr"), "PAGE")
     footer._p.append(fld)
 
+    # Save
     bio = io.BytesIO()
     doc.save(bio)
     bio.seek(0)
@@ -128,7 +134,7 @@ def create_charts(data):
     obs = float(data.get("observation_average", 0))
     fb = float(data.get("feedback_average", 0))
     fig, ax = plt.subplots(figsize=(5, 3))
-    ax.bar(["Observation", "Feedback"], [obs, fb], color=["#5DADE2", "#48C9B0"])
+    ax.bar(["Observation", "Feedback"], [obs, fb])
     ax.set_ylim(0, 5)
     ax.set_ylabel("Score (1-5)")
     plt.tight_layout()
@@ -136,81 +142,41 @@ def create_charts(data):
     fig.savefig(buf, format="png")
     plt.close(fig)
     charts.append((buf.getvalue(), "Observation vs Feedback (1–5 scale)"))
-
-    teachers = int(data.get("total_teachers", 0))
-    nonteach = int(data.get("total_non_teaching_staff", 0))
-    fig2, ax2 = plt.subplots()
-    ax2.pie([teachers, nonteach], labels=["Teachers", "Non-Teaching"], autopct="%1.0f%%")
-    plt.tight_layout()
-    buf2 = io.BytesIO()
-    fig2.savefig(buf2, format="png")
-    plt.close(fig2)
-    charts.append((buf2.getvalue(), "Staff Composition"))
     return charts
 
 
-def build_urdu_md(report):
-    lines = report.split("\n")
-    text = "<div dir='rtl' style='line-height:1.5;'>\n\n"
-    for line in lines:
-        if not line.strip():
-            continue
-        if line.strip().endswith("۔") or line.strip().endswith(":"):
-            text += f"<b>{line.strip()}</b>\n\n"
-        else:
-            text += line.strip() + "\n\n"
-    text += "</div>"
-    return text.encode("utf-8")
+def sample_report(data):
+    name = data.get("name", "School")
+    return (
+        f"Executive Summary:\n{name} demonstrates strong performance.\n\n"
+        "Domain-wise Findings:\nTeaching quality and management are satisfactory.\n\n"
+        "Key Findings Table:\nInfrastructure needs improvement in labs.\n\n"
+        "Recommendations and SMART Action Plan:\n"
+        "1. Conduct monthly assessments.\n2. Improve classroom facilities.\n\n"
+        "Overall Audit Rating (1–5): 4.2\n\n"
+        "Conclusion:\nThe institution is performing effectively.\n"
+    )
 
 
-def sample_report(data, language):
-    name = data.get("name", "School")
-    if language == "Urdu":
-        return (
-            f"خلاصہ:\n{name} ایک تعلیمی ادارہ ہے جو مسلسل بہتری کی سمت میں گامزن ہے۔\n\n"
-            "مشاہداتی اوسط: 4.2\nفیڈبیک اوسط: 4.0\n\n"
-            "تجاویز:\n1. اساتذہ کی تربیت میں اضافہ کیا جائے۔\n2. انفراسٹرکچر میں بہتری لائی جائے۔\n\n"
-            "نتیجہ:\nادارہ مجموعی طور پر اچھی کارکردگی دکھا رہا ہے۔\n\n"
-            "Developed by Najaf Ali Sharqi"
-        )
-    else:
-        return (
-            f"Executive Summary:\n{name} demonstrates strong educational performance.\n\n"
-            "Findings:\nTeaching quality and management are satisfactory.\n\n"
-            "Recommendations:\n1. Conduct monthly internal assessments.\n2. Improve infrastructure.\n\n"
-            "Conclusion:\nThe institution is progressing effectively.\n\n"
-            "Developed by Najaf Ali Sharqi"
-        )
-
-
-def generate_report_from_data(school_data, language):
-    prompt = build_prompt(school_data, language)
+def generate_report_from_data(school_data):
+    """Main report generation function."""
+    prompt = build_prompt(school_data)
     try:
-        report = generate_with_gemini(prompt)
+        report = generate_with_groq(prompt)
         report = sanitize_text(report)
     except Exception:
-        report = sample_report(school_data, language)
-
-    charts = create_charts(school_data) if language == "English" else []
-    if language == "English":
-        file_bytes = create_docx(report, school_data, charts)
-        fname = f"EPA_Report_{school_data.get('name','school').replace(' ','_')}.docx"
-    else:
-        file_bytes = build_urdu_md(report)
-        fname = f"EPA_Report_{school_data.get('name','school').replace(' ','_')}.md"
+        report = sample_report(school_data)
+
+    charts = create_charts(school_data)
+    file_bytes = create_docx(report, school_data, charts)
+    fname = f"EPA_Report_{school_data.get('name','school').replace(' ','_')}.docx"
     return report, fname, file_bytes
 
 
-# === GRADIO UI ===
+# === GRADIO INTERFACE ===
 with gr.Blocks(title="Educational Pre-Audit (EPA) Generator") as demo:
     gr.Markdown("## 🏫 Educational Pre-Audit (EPA) Generator\n**Developed by Najaf Ali Sharqi**")
 
-    language = gr.Dropdown(["English", "Urdu"], value="English", label="Select Language")
-
-    file_input = gr.File(label="Upload School Data File (.json or .csv)", file_types=[".json", ".csv"])
-
-    gr.Markdown("### Manual Data Entry")
-
     with gr.Row():
         name = gr.Textbox(label="Institution Name", value="Highlanders Public School and College, Skardu")
         principal = gr.Textbox(label="Principal Name", value="Mr. Karim Ali")
@@ -232,27 +198,16 @@ with gr.Blocks(title="Educational Pre-Audit (EPA) Generator") as demo:
     policy_gaps = gr.Textbox(label="Policy Gaps", value="HR policy missing; incomplete reports", lines=3)
     results_summary = gr.Textbox(label="Results Summary", value="Pass Rate 92%; GPA 3.4", lines=2)
 
-    generate_btn = gr.Button("🚀 Generate Report")
+    generate_btn = gr.Button("🚀 Generate English Report")
     preview = gr.Textbox(label="Preview", lines=20)
     download = gr.File(label="Download Report")
 
     def on_generate(
-        language, file_input, name, principal, location, level, total_students, total_teachers,
-        total_non_teaching_staff, observation_average, feedback_average, infrastructure, policy_gaps, results_summary
+        name, principal, location, level, total_students, total_teachers,
+        total_non_teaching_staff, observation_average, feedback_average,
+        infrastructure, policy_gaps, results_summary
     ):
-        data = {}
-        if file_input is not None:
-            ext = os.path.splitext(file_input.name)[1].lower()
-            try:
-                if ext == ".json":
-                    data = json.load(open(file_input.name, "r", encoding="utf-8"))
-                elif ext == ".csv":
-                    df = pd.read_csv(file_input.name)
-                    data = df.to_dict(orient="records")[0]
-            except Exception:
-                pass
-
-        data.update({
+        data = {
             "name": name,
             "principal": principal,
             "location": location,
@@ -265,9 +220,9 @@ with gr.Blocks(title="Educational Pre-Audit (EPA) Generator") as demo:
             "infrastructure": infrastructure,
             "policy_gaps": policy_gaps,
             "results": results_summary,
-        })
+        }
 
-        report, fname, fbytes = generate_report_from_data(data, language)
+        report, fname, fbytes = generate_report_from_data(data)
         tmp = f"/tmp/{fname}"
         with open(tmp, "wb") as f:
             f.write(fbytes)
@@ -275,10 +230,10 @@ with gr.Blocks(title="Educational Pre-Audit (EPA) Generator") as demo:
 
     generate_btn.click(
         on_generate,
-        inputs=[language, file_input, name, principal, location, level, total_students, total_teachers,
-                total_non_teaching_staff, observation_average, feedback_average, infrastructure, policy_gaps,
-                results_summary],
-        outputs=[preview, download]
+        inputs=[name, principal, location, level, total_students, total_teachers,
+                total_non_teaching_staff, observation_average, feedback_average,
+                infrastructure, policy_gaps, results_summary],
+        outputs=[preview, download],
     )
 
 if __name__ == "__main__":