Rudimentary helper to check that your dbt documentation matches your model

dbt_check_docs_match_code.py

#!/usr/bin/env python3
import json
import os

import pandas as pd

abs_path = os.path.dirname(os.path.abspath(__file__ + '/../'))
catalog_path = os.path.join(abs_path, 'target/catalog.json')
manifest_path = os.path.join(abs_path, 'target/manifest.json')

catalog_df = pd.read_json(catalog_path)
data = json.load(open(manifest_path))

manifest_df = pd.DataFrame(data["nodes"]).T

catalog_dict = {}
for index, row in catalog_df.iterrows():
    catalog_dict[row['nodes']['metadata']['name']] = row['nodes']['columns'].keys()
for index, row in manifest_df.iterrows():
    if row['resource_type'] == 'model':
        name = row['name']
        manifest_columns = set(row['columns'].keys())
        try:
            catalog_columns = set(catalog_dict[name])

            if len(manifest_columns.symmetric_difference(catalog_columns)) != 0:
                print("\n")
                print(name)
                manifest_difference = manifest_columns - catalog_columns
                catalog_difference = catalog_columns - manifest_columns
                if len(manifest_difference) != 0:
                    print("In docs but not in sql:")
                    print(manifest_difference)
                if len(catalog_difference) != 0:
                    print("In sql but not in docs:")
                    print(catalog_difference)
        except:
            print(f"{name} failed to parse")