Add app.py

app.py

@@ -0,0 +1,159 @@
+"""
+VynFi Streamlit Template — Generate, Explore, Visualize
+
+A ready-to-deploy Streamlit app that connects to the VynFi API,
+generates synthetic financial data, and renders interactive dashboards.
+Clone and customize for your use case.
+
+Usage:
+    export VYNFI_API_KEY=vf_live_...
+    streamlit run app.py
+"""
+
+import os
+import streamlit as st
+import pandas as pd
+import vynfi
+
+st.set_page_config(page_title="VynFi Explorer", page_icon="📊", layout="wide")
+
+st.title("📊 VynFi Data Explorer")
+
+api_key = os.environ.get("VYNFI_API_KEY", "")
+if not api_key:
+    api_key = st.sidebar.text_input("VynFi API Key", type="password", placeholder="vf_live_...")
+
+if not api_key:
+    st.info("Enter your VynFi API key in the sidebar to get started. [Get a free key →](https://vynfi.com/signup)")
+    st.stop()
+
+client = vynfi.VynFi(api_key=api_key)
+
+# ── Sidebar: generation config ──────────────────────────────────────────────
+
+st.sidebar.header("Generate")
+
+sector = st.sidebar.selectbox(
+    "Sector",
+    ["retail", "manufacturing", "financial_services", "banking_aml", "healthcare", "technology", "energy"],
+    index=1,
+)
+rows = st.sidebar.slider("Rows", min_value=100, max_value=100_000, value=1000, step=100)
+companies = st.sidebar.slider("Companies", min_value=1, max_value=20, value=3)
+fraud_rate = st.sidebar.slider("Fraud rate", min_value=0.0, max_value=0.20, value=0.03, step=0.01)
+
+# NL config option
+st.sidebar.divider()
+nl_description = st.sidebar.text_area(
+    "Or describe what you want (Scale+)",
+    placeholder="e.g. 6 months of P2P for a German manufacturer with IFRS",
+    height=80,
+)
+
+generate = st.sidebar.button("Generate", type="primary", use_container_width=True)
+
+# ── Main area ────────────────────────────────────────────────────────────────
+
+if generate:
+    with st.spinner("Generating..."):
+        try:
+            if nl_description.strip():
+                resp = client._request("POST", "/v1/configs/from-description", json={
+                    "description": nl_description.strip()
+                })
+                config = resp.get("config", {})
+                st.sidebar.success(f"AI config: {config.get('sector')} / {config.get('rows')} rows")
+            else:
+                config = {
+                    "sector": sector,
+                    "rows": rows,
+                    "companies": companies,
+                    "fraudRate": fraud_rate,
+                    "complexity": "medium",
+                    "exportFormat": "json",
+                    "output": {"numericMode": "native"},
+                }
+
+            job = client.jobs.generate_config(config=config)
+            completed = client.jobs.wait(job.id, poll_interval=3.0, timeout=300.0)
+
+            if completed.status != "completed":
+                st.error(f"Job failed: {completed.error_detail}")
+                st.stop()
+
+            st.session_state["job_id"] = completed.id
+            st.session_state["archive"] = client.jobs.download_archive(completed.id)
+            st.success(f"Job {completed.id} completed")
+        except Exception as e:
+            st.error(f"Error: {e}")
+
+# ── Display results ──────────────────────────────────────────────────────────
+
+if "archive" in st.session_state:
+    archive = st.session_state["archive"]
+
+    tab1, tab2, tab3, tab4 = st.tabs(["Journal Entries", "Documents", "Quality", "Files"])
+
+    with tab1:
+        st.subheader("Journal Entries")
+        try:
+            entries = archive.json("journal_entries.json")
+            # Flatten
+            rows_flat = []
+            for entry in entries[:500]:  # Cap for display
+                header = entry.get("header", entry)
+                lines = entry.get("lines", [entry])
+                for line in lines:
+                    rows_flat.append({
+                        "document_id": header.get("document_id", ""),
+                        "company_code": header.get("company_code", ""),
+                        "posting_date": header.get("posting_date", ""),
+                        "document_type": header.get("document_type", ""),
+                        "is_fraud": header.get("is_fraud", False),
+                        "gl_account": line.get("gl_account", ""),
+                        "debit_amount": line.get("debit_amount", 0),
+                        "credit_amount": line.get("credit_amount", 0),
+                    })
+            df = pd.DataFrame(rows_flat)
+            st.metric("Line items", f"{len(df):,}")
+
+            col1, col2 = st.columns(2)
+            with col1:
+                st.metric("Total debits", f"${df['debit_amount'].sum():,.2f}")
+            with col2:
+                fraud_count = df["is_fraud"].sum()
+                st.metric("Fraud entries", f"{fraud_count} ({fraud_count/len(df)*100:.1f}%)")
+
+            st.dataframe(df, use_container_width=True, hide_index=True)
+        except Exception as e:
+            st.warning(f"Could not load journal entries: {e}")
+
+    with tab2:
+        st.subheader("Document Flows")
+        for doc_type in ["purchase_orders", "goods_receipts", "vendor_invoices", "payments"]:
+            try:
+                docs = archive.json(f"document_flows/{doc_type}.json")
+                st.write(f"**{doc_type.replace('_', ' ').title()}**: {len(docs)} records")
+            except Exception:
+                pass
+
+    with tab3:
+        st.subheader("Quality Metrics")
+        try:
+            analytics = client.jobs.analytics(st.session_state["job_id"])
+            if hasattr(analytics, "benford_analysis") and analytics.benford_analysis:
+                b = analytics.benford_analysis
+                col1, col2, col3 = st.columns(3)
+                col1.metric("Benford MAD", f"{b.mad:.4f}")
+                col2.metric("Chi-squared", f"{b.chi_squared:.2f}")
+                col3.metric("Conforms", "✅" if b.passes else "❌")
+        except Exception:
+            st.info("Quality analytics not available for this job.")
+
+    with tab4:
+        st.subheader("Archive Files")
+        for f in archive.files():
+            st.text(f)
+
+else:
+    st.info("Click **Generate** in the sidebar to create a dataset.")