mdbt 0.4.38__tar.gz → 0.4.39__tar.gz

{mdbt-0.4.38 → mdbt-0.4.39}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mdbt
-Version: 0.4.38
+Version: 0.4.39
 Summary: A CLI tool to manage dbt builds with state handling and manifest management
 Author: Craig Lathrop
 Author-email: info@markimicrowave.com

{mdbt-0.4.38 → mdbt-0.4.39}/mdbt/build_dbt_docs_ai.py

@@ -1,8 +1,10 @@
 import subprocess
+from pathlib import Path
 
 import pyperclip
 from dotenv import find_dotenv
 from dotenv import load_dotenv
+from ruamel.yaml import YAML
 
 from mdbt.ai_core import AiCore
 from mdbt.prompts import Prompts
@@ -50,8 +52,13 @@ class BuildDBTDocs(AiCore):
 
             system_instructions = system_instructions + sample_data[model_name]
 
+            # Get existing column descriptions from neighboring YML files
+            existing_descriptions = self.get_existing_column_descriptions(original_file_path)
+            if existing_descriptions:
+                system_instructions += f'Here are some existing column descriptions from neighboring models. These may or may not be relevant to the current model. Assess, and use as a source if it seems to match the sample data: \n{existing_descriptions}'
+
             # Might bring this back in the future.
-            extra_info = ""
+            extra_info = ''
 
             if mode == 1:
                 # Build new documentation
@@ -113,11 +120,94 @@ class BuildDBTDocs(AiCore):
 
     @staticmethod
     def backup_existing_yml_file(yml_file_path):
-        with open(yml_file_path, "r") as file:
+        with open(yml_file_path, 'r') as file:
             yml_content = file.read()
-        with open(yml_file_path + ".bak", "w") as file:
+        with open(yml_file_path + '.bak', 'w') as file:
             file.write(yml_content)
 
+    def get_existing_column_descriptions(self, model_file_path: str) -> str:
+        """
+        Search for YML files in the same folder as the model and extract
+        existing column descriptions to provide context for consistent naming.
+
+        Args:
+            model_file_path: Path to the SQL model file.
+
+        Returns:
+            A formatted string containing existing column descriptions found
+            in neighboring YML files.
+        """
+        folder_path = Path(model_file_path).parent
+        current_yml_path = model_file_path.replace('.sql', '.yml')
+        column_descriptions = {}
+
+        yaml = YAML()
+        yaml.preserve_quotes = True
+
+        # Find all .yml files in the same folder
+        yml_files = list(folder_path.glob('*.yml'))
+
+        for yml_file in yml_files:
+            # Skip the current model's yml file if it exists
+            if str(yml_file) == current_yml_path:
+                continue
+
+            try:
+                with open(yml_file, 'r') as f:
+                    yml_content = yaml.load(f)
+
+                if not yml_content:
+                    continue
+
+                # Extract column descriptions from models section
+                models = yml_content.get('models', [])
+                for model in models:
+                    columns = model.get('columns', [])
+                    for column in columns:
+                        col_name = column.get('name')
+                        col_desc = column.get('description')
+                        if col_name and col_desc:
+                            # Only store if not already seen, or prefer longer descriptions
+                            if col_name not in column_descriptions or \
+                               len(str(col_desc)) > len(str(column_descriptions[col_name])):
+                                column_descriptions[col_name] = str(col_desc).strip()
+
+            except Exception as e:
+                print(f'Warning: Could not parse {yml_file}: {e}')
+                continue
+
+        if not column_descriptions:
+            return ''
+
+        # Filter out ID columns to reduce noise (columns ending in _id or named id)
+        non_id_descriptions = {
+            col_name: desc for col_name, desc in column_descriptions.items()
+            if not col_name.lower().endswith('_id') and col_name.lower() != 'id'
+        }
+
+        # Limit to first 300 unique columns to avoid oversized prompts
+        max_columns = 300
+        limited_descriptions = dict(
+            list(sorted(non_id_descriptions.items()))[:max_columns]
+        )
+
+        if not limited_descriptions:
+            return ''
+
+        # Format the output
+        lines = [
+            '\n--- Existing Column Descriptions from Neighboring Models ---',
+            'Use these existing descriptions for consistency when the same '
+            'or similar columns appear in the new model:\n'
+        ]
+        for col_name, description in limited_descriptions.items():
+            # Truncate very long descriptions for context
+            desc_preview = description[:200] + '...' if len(description) > 200 else description
+            lines.append(f'  - {col_name}: {desc_preview}')
+
+        lines.append('--- End of Existing Column Descriptions ---\n')
+        return '\n'.join(lines)
+
     def build_user_msg_mode_1(self, _sql_file_path: str, extra_info: str) -> str:
         self.read_file(_sql_file_path)
         model_name = _sql_file_path.split("/")[-1].split(".")[0]

{mdbt-0.4.38 → mdbt-0.4.39}/mdbt.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mdbt
-Version: 0.4.38
+Version: 0.4.39
 Summary: A CLI tool to manage dbt builds with state handling and manifest management
 Author: Craig Lathrop
 Author-email: info@markimicrowave.com

{mdbt-0.4.38 → mdbt-0.4.39}/setup.py

@@ -3,7 +3,7 @@ from setuptools import setup
 
 setup(
     name="mdbt",
-    version="0.4.38",
+    version="0.4.39",
     description="A CLI tool to manage dbt builds with state handling and manifest management",
     author="Craig Lathrop",
     author_email="info@markimicrowave.com",