Fix all mypy type errors in CI-failing files

- Add type annotations for results dictionaries in proc_print, proc_sql, proc_language
- Add None checks for self.generator in proc_language methods
- Fix notebook_interpreter return types with type: ignore comments
- Fix kernel signature mismatches by adding missing parameters
- Add Optional type annotation for self.interpreter in statlang_kernel
- Add type guards for all self.interpreter usages

All originally failing files now pass mypy type checking.
All tests still pass.

Author: Ryan Story

stat_lang/kernel/statlang_kernel.py

@@ -60,6 +60,7 @@ def __init__(self, **kwargs):
         logger.info("=" * 80)
 
         super().__init__(**kwargs)
+        self.interpreter: Optional[SASInterpreter] = None
         try:
             self.interpreter = SASInterpreter()
             # Kernel base class initializes execution_count, but ensure it starts at 0
@@ -76,7 +77,7 @@ def __init__(self, **kwargs):
             print(f"Warning: Failed to initialize interpreter: {e}")
             self.interpreter = None
 
-    def do_execute(self, code, silent, store_history=True, user_expressions=None, allow_stdin=False):
+    def do_execute(self, code, silent, store_history=True, user_expressions=None, allow_stdin=False, *, cell_meta=None, cell_id=None):
         """Execute StatLang code in the kernel."""
 
         logger.info(f"do_execute called with code: {repr(code[:100])}...")
@@ -135,6 +136,8 @@ def do_execute(self, code, silent, store_history=True, user_expressions=None, al
         self.error_buffer = io.StringIO()
 
         # Record datasets before execution
+        if self.interpreter is None:
+            return {'status': 'error', 'ename': 'InterpreterError', 'evalue': 'Interpreter not initialized'}
         datasets_before = set(self.interpreter.data_sets.keys())
         logger.info(f"Datasets before execution: {datasets_before}")
 
@@ -176,7 +179,8 @@ def do_execute(self, code, silent, store_history=True, user_expressions=None, al
                     self._send_datasets_display(datasets)
                 else:
                     # Reset the flag for next execution
-                    self.interpreter._suppress_dataset_display = False
+                    if self.interpreter is not None:
+                        self.interpreter._suppress_dataset_display = False
 
             logger.info("Returning successful execution result")
             # Increment execution count AFTER successful execution
@@ -250,7 +254,7 @@ def do_complete(self, code, cursor_pos):
             'status': 'ok'
         }
 
-    def do_inspect(self, code, cursor_pos, detail_level=0):
+    def do_inspect(self, code, cursor_pos, detail_level=0, omit_sections=()):
         """Provide code inspection/hover information."""
         # Get the word at cursor position
         text_before_cursor = code[:cursor_pos]
@@ -281,6 +285,8 @@ def do_inspect(self, code, cursor_pos, detail_level=0):
 
     def _get_datasets_info(self):
         """Get information about datasets created in the interpreter."""
+        if self.interpreter is None:
+            return {}
         datasets = {}
         for name, df in self.interpreter.data_sets.items():
             datasets[name] = {
@@ -294,6 +300,8 @@ def _get_datasets_info(self):
 
     def _get_new_datasets_info(self, datasets_before):
         """Get information about datasets created in the current execution."""
+        if self.interpreter is None:
+            return {}
         datasets = {}
         current_datasets = set(self.interpreter.data_sets.keys())
         new_datasets = current_datasets - datasets_before

stat_lang/kernel/working_kernel.py

@@ -35,7 +35,7 @@ def __init__(self, **kwargs):
         self.error_buffer = io.StringIO()
         self.datasets_before_execution = set()
 
-    def do_execute(self, code, silent, store_history=True, user_expressions=None, allow_stdin=False):
+    def do_execute(self, code, silent, store_history=True, user_expressions=None, allow_stdin=False, *, cell_meta=None, cell_id=None):
         """Execute StatLang code in the kernel."""
 
         # Skip empty cells
@@ -151,7 +151,7 @@ def do_complete(self, code, cursor_pos):
             'status': 'ok'
         }
 
-    def do_inspect(self, code, cursor_pos, detail_level=0):
+    def do_inspect(self, code, cursor_pos, detail_level=0, omit_sections=()):
         """Provide code inspection/hover information."""
         # Get the word at cursor position
         text_before_cursor = code[:cursor_pos]

stat_lang/notebook_interpreter.py

@@ -23,7 +23,7 @@ def __init__(self):
         self.output_buffer = io.StringIO()
         self.error_buffer = io.StringIO()
 
-    def run_code(self, sas_code: str) -> Dict[str, Any]:
+    def run_code(self, sas_code: str) -> Dict[str, Any]:  # type: ignore[override]
         """
         Run SAS code and return structured results for notebook display.
         
@@ -91,7 +91,7 @@ def _get_summary_stats(self, df: pd.DataFrame) -> Dict[str, Any]:
         if len(numeric_cols) == 0:
             return {}
 
-        return df[numeric_cols].describe().to_dict()
+        return df[numeric_cols].describe().to_dict()  # type: ignore[no-any-return]
 
     def _get_proc_results(self) -> List[Dict[str, Any]]:
         """Get results from PROC procedures executed in this session."""
@@ -168,11 +168,11 @@ def export_dataset(self, name: str, format: str = 'json') -> str:
             return f"Dataset {name} not found"
 
         if format == 'json':
-            return df.to_json(orient='records', indent=2)
+            return str(df.to_json(orient='records', indent=2))
         elif format == 'csv':
-            return df.to_csv(index=False)
+            return str(df.to_csv(index=False))
         elif format == 'html':
-            return df.to_html(index=False, classes='table table-striped')
+            return str(df.to_html(index=False, classes='table table-striped'))
         else:
             return f"Unsupported format: {format}"
 

stat_lang/procs/proc_language.py

@@ -57,7 +57,7 @@ def execute(self, data: pd.DataFrame, proc_info: ProcStatement, dataset_manager=
         Returns:
             Dictionary containing results and output data
         """
-        results = {
+        results: Dict[str, Any] = {
             'output_text': [],
             'output_data': None
         }
@@ -122,6 +122,11 @@ def execute(self, data: pd.DataFrame, proc_info: ProcStatement, dataset_manager=
 
     def _generate_text(self, prompt: str, model: str) -> Dict[str, Any]:
         """Generate text using the Hugging Face model."""
+        if self.generator is None:
+            return {
+                'output': ["ERROR: Language model not initialized."],
+                'data': None
+            }
         try:
             # Generate text using the pipeline
             result = self.generator(
@@ -161,6 +166,11 @@ def _generate_text(self, prompt: str, model: str) -> Dict[str, Any]:
 
     def _question_answer(self, question: str, context: str, model: str) -> Dict[str, Any]:
         """Answer questions using the model."""
+        if self.generator is None:
+            return {
+                'output': ["ERROR: Language model not initialized."],
+                'data': None
+            }
         try:
             # For Q&A, we'll use text generation with a structured prompt
             if context:
@@ -208,6 +218,11 @@ def _question_answer(self, question: str, context: str, model: str) -> Dict[str,
 
     def _summarize_text(self, text: str, model: str) -> Dict[str, Any]:
         """Summarize text using the model."""
+        if self.generator is None:
+            return {
+                'output': ["ERROR: Language model not initialized."],
+                'data': None
+            }
         try:
             summary_prompt = f"Summarize the following text:\n\n{text}\n\nSummary:"
 
@@ -249,6 +264,11 @@ def _summarize_text(self, text: str, model: str) -> Dict[str, Any]:
 
     def _analyze_text(self, text: str, model: str) -> Dict[str, Any]:
         """Analyze text using the model."""
+        if self.generator is None:
+            return {
+                'output': ["ERROR: Language model not initialized."],
+                'data': None
+            }
         try:
             analysis_prompt = f"Analyze the following text and provide insights:\n\n{text}\n\nAnalysis:"
 
@@ -290,6 +310,11 @@ def _analyze_text(self, text: str, model: str) -> Dict[str, Any]:
 
     def _summarize_data(self, data: pd.DataFrame, var_vars: List[str], prompt: str, model: str) -> Dict[str, Any]:
         """Summarize data using the model."""
+        if self.generator is None:
+            return {
+                'output': ["ERROR: Language model not initialized."],
+                'data': None
+            }
         try:
             # Create a text description of the data
             data_desc = f"Dataset with {len(data)} rows and {len(data.columns)} columns."
@@ -337,6 +362,11 @@ def _summarize_data(self, data: pd.DataFrame, var_vars: List[str], prompt: str,
 
     def _analyze_data(self, data: pd.DataFrame, var_vars: List[str], prompt: str, model: str) -> Dict[str, Any]:
         """Analyze data using the model."""
+        if self.generator is None:
+            return {
+                'output': ["ERROR: Language model not initialized."],
+                'data': None
+            }
         try:
             # Create a text description of the data
             data_desc = f"Dataset with {len(data)} rows and {len(data.columns)} columns."

stat_lang/procs/proc_print.py

@@ -31,7 +31,7 @@ def execute(self, data: pd.DataFrame, proc_info: ProcStatement, dataset_manager=
         Returns:
             Dictionary containing results and output data
         """
-        results = {
+        results: Dict[str, Any] = {
             'output_text': [],
             'output_data': None
         }

stat_lang/procs/proc_sql.py

@@ -40,7 +40,7 @@ def execute(self, data: pd.DataFrame, proc_info: ProcStatement, dataset_manager=
         Returns:
             Dictionary containing results and output data
         """
-        results = {
+        results: Dict[str, Any] = {
             'output_text': [],
             'output_data': None
         }