update sql_generator notebook: schema-only mode, all 12 pipelines

- Switch from DuckDB file to schema-only transpilation (no data needed)
- Add cclf pipeline (previously excluded due to transpile errors)
- Include cclf in summary table
- Remove try/except wrapper and hardcoded DB path

notebooks/sql_generator.py

@@ -20,21 +20,17 @@ def _(mo):
     narwhals express functions against DuckDB relations and extracting the SQL
     via `.sql_query()`, then transpiling with sqlglot.
 
-    Each cell below transpiles one pipeline module and displays the generated
-    `CREATE OR REPLACE TABLE` statements.
+    Uses schema-only DuckDB (no data file needed). Each cell below transpiles
+    one pipeline module and displays the generated `CREATE OR REPLACE TABLE`
+    statements.
     """)
     return
 
 
 @app.cell(hide_code=True)
 def _():
-    import duckdb
-
     from aco.lake.transpile import transpile
 
-    DB_PATH = "/home/kert/notebooks/aco.duckdb"
-    con = duckdb.connect(DB_PATH, read_only=True)
-
     DIALECT = "databricks"
     CATALOG = "homelab"
 
@@ -42,7 +38,6 @@ def _():
         """Transpile a pipeline and return formatted SQL."""
         result = transpile(
             pipeline,
-            con,
             target_dialect=DIALECT,
             catalog=CATALOG,
             output_mode="ctas",
@@ -56,7 +51,7 @@ def _():
         header = f"/* {label}: {len(result)} steps, {errors} errors */\n"
         return header + "\n\n".join(parts), len(result), errors
 
-    return con, run_transpile
+    return (run_transpile,)
 
 
 @app.cell(hide_code=True)
@@ -241,13 +236,8 @@ def _(mo):
 def _(mo, run_transpile):
     from aco.pipe import cclf as _cclf
 
-    try:
-        _sql, _n, _e = run_transpile(_cclf.pipeline, "cclf")
-        _msg = f"**{_n} steps, {_e} errors**\n\n```sql\n{_sql}\n```"
-    except Exception as _ex:
-        _msg = f"**CCLF pipeline requires CCLF data in DuckDB.**\n\nError: `{_ex}`"
-
-    mo.md(_msg)
+    _sql, _n, _e = run_transpile(_cclf.pipeline, "cclf")
+    mo.md(f"**{_n} steps, {_e} errors**\n\n```sql\n{_sql}\n```")
     return
 
 
@@ -278,10 +268,10 @@ def _(mo):
 
 
 @app.cell
-def _(con, mo):
-    from aco.lake.transpile import transpile as _transpile
+def _(mo, run_transpile):
     from aco.pipe import (
         ahrq_measures,
+        cclf,
         claims_preprocessing,
         core,
         data_quality,
@@ -305,6 +295,7 @@ def _(con, mo):
         ("quality_measures", quality_measures),
         ("hcc_suspecting", hcc_suspecting),
         ("provider_attribution", provider_attribution),
+        ("cclf", cclf),
         ("main", main),
     ]
 
@@ -312,13 +303,10 @@ def _(con, mo):
     _total_steps = 0
     _total_errors = 0
     for _name, _mod in _modules:
-        _result = _transpile(
-            _mod.pipeline, con, target_dialect="databricks", catalog="homelab"
-        )
-        _errs = sum(1 for v in _result.values() if v.startswith("-- ERROR"))
-        _total_steps += len(_result)
-        _total_errors += _errs
-        _rows.append(f"| {_name} | {len(_result)} | {_errs} |")
+        _, _n, _e = run_transpile(_mod.pipeline, _name)
+        _total_steps += _n
+        _total_errors += _e
+        _rows.append(f"| {_name} | {_n} | {_e} |")
 
     _table = "| Pipeline | Steps | Errors |\n|----------|-------|--------|\n"
     _table += "\n".join(_rows)