fix: ui.py — force PYTHONIOENCODING=utf-8 and NO_COLOR=1 in subprocess to prevent Windows cp1252 crash"

alpha_factory/ui.py

@@ -11,7 +11,6 @@ import duckdb
 import gradio as gr
 from pathlib import Path
 
-# Load .env so HF_TOKEN is available
 try:
     from dotenv import load_dotenv
     load_dotenv()
@@ -52,7 +51,7 @@ def get_alphas_from_db(limit=50):
 def get_alpha_cards():
     rows = get_alphas_from_db()
     if not rows:
-        return [["No alphas yet", "", "", "", "", "", "", "Run a batch first"]]
+        return [["—", "—", "—", "—", "—", "—", "—", "Run a batch first"]]
     data = []
     for row in rows:
         alpha_id, submitted_at, expression, theme, archetype, tag, neutral, decay, fields, verdict = row
@@ -74,9 +73,15 @@ def get_full_expression(evt: gr.SelectData):
 
 
 def run_batch(batch_size):
-    """Run pipeline as subprocess and return log."""
+    """Run pipeline as subprocess with forced UTF-8 to avoid Windows encoding crash."""
     env = os.environ.copy()
-    # Ensure HF_TOKEN passes to subprocess
+    # Force UTF-8 output — prevents Rich/Windows cp1252 crash
+    env["PYTHONIOENCODING"] = "utf-8"
+    env["PYTHONLEGACYWINDOWSSTDIO"] = "utf-8"
+    # Disable Rich color/formatting when piped (cleaner output)
+    env["NO_COLOR"] = "1"
+    env["TERM"] = "dumb"
+    # Ensure HF_TOKEN passes through
     if "HF_TOKEN" not in env:
         token = os.getenv("HF_TOKEN", "")
         if token:
@@ -86,29 +91,29 @@ def run_batch(batch_size):
         result = subprocess.run(
             [sys.executable, "-m", "alpha_factory.run", "--dry-run", "--batch-size", str(int(batch_size))],
             capture_output=True,
-            text=True,
-            encoding="utf-8",
-            errors="replace",
             env=env,
-            timeout=120,
+            timeout=180,
             cwd=str(Path.cwd()),
         )
+        # Decode with utf-8, replace errors
+        stdout = result.stdout.decode("utf-8", errors="replace") if result.stdout else ""
+        stderr = result.stderr.decode("utf-8", errors="replace") if result.stderr else ""
+        
         log = ""
-        if result.stdout:
-            log += result.stdout[-3000:]
-        if result.returncode != 0 and result.stderr:
-            log += "\n\n--- ERRORS ---\n" + result.stderr[-2000:]
+        if stdout:
+            log = stdout[-3000:]
+        if result.returncode != 0 and stderr:
+            log += "\n\n--- ERRORS ---\n" + stderr[-2000:]
         if not log.strip():
             log = f"Process exited with code {result.returncode}"
         return log
     except subprocess.TimeoutExpired:
-        return "ERROR: Pipeline timed out after 120 seconds"
+        return "ERROR: Pipeline timed out after 180 seconds. Try smaller batch size."
     except Exception as e:
         return f"ERROR: {str(e)}"
 
 
 def generate_and_refresh(batch_size):
-    """Run batch then refresh the table."""
     log = run_batch(batch_size)
     table = get_alpha_cards()
     return table, log
@@ -117,20 +122,20 @@ def generate_and_refresh(batch_size):
 def build_ui():
     with gr.Blocks(title="Alpha Factory") as app:
         gr.Markdown("""
-        # 🏭 Alpha Factory — Generated Alphas
+        # Alpha Factory — Generated Alphas
         View, copy, and manage alphas generated by the pipeline.
         """)
 
         with gr.Row():
             with gr.Column(scale=1):
                 batch_size_input = gr.Number(value=3, label="Batch Size", minimum=1, maximum=20)
-                generate_btn = gr.Button("🚀 Generate New Batch", variant="primary")
-                refresh_btn = gr.Button("🔄 Refresh Table")
+                generate_btn = gr.Button("Generate New Batch", variant="primary")
+                refresh_btn = gr.Button("Refresh Table")
                 gr.Markdown("*Dry run mode — no BRAIN submissions*")
             with gr.Column(scale=3):
                 stats_md = gr.Markdown(f"**Alphas in store:** {len(get_alphas_from_db())}")
 
-        gr.Markdown("### 📋 Click any row to see full expression")
+        gr.Markdown("### Click any row to see full expression")
 
         alpha_table = gr.Dataframe(
             value=get_alpha_cards(),
@@ -139,14 +144,14 @@ def build_ui():
             wrap=True,
         )
 
-        gr.Markdown("### 📝 Full Expression — Ctrl+A then Ctrl+C to copy")
+        gr.Markdown("### Full Expression — Ctrl+A then Ctrl+C to copy")
         full_expr = gr.Textbox(
             label="Full Expression",
             lines=6,
-            interactive=True,  # So user can select & copy
+            interactive=True,
         )
 
-        gr.Markdown("### 📜 Pipeline Log")
+        gr.Markdown("### Pipeline Log")
         pipeline_log = gr.Textbox(label="Output", lines=15, interactive=False)
 
         # Events