Skip to content

Instantly share code, notes, and snippets.

@ducchetrongminh
Last active January 29, 2024 20:57
Show Gist options
  • Save ducchetrongminh/c494d867feec925a5b59515714778279 to your computer and use it in GitHub Desktop.
Save ducchetrongminh/c494d867feec925a5b59515714778279 to your computer and use it in GitHub Desktop.
Python script to auto generating dbt docs block from yml files
# Standard imports
import os
# Library imports
import yaml
# Local imports
"""
This script will generate dbt docs block from documentation inside yml files. This allows us to reuse existing dbt documentation.
Use official solution when this issue is fixed: https://github.com/dbt-labs/dbt-core/issues/2995.
To generate docs block files:
python -m path.to.dbt_docsblock_autogenerator
It will generate these files:
- 1st-docs-path/autogen_models_docsblock.md
- 1st-docs-path/autogen_sources_docsblock.md
To reuse documentation, use docs block with the following format:
- For model:
+ autogen__model_name
+ autogen__model_name__column_name
- For source:
+ autogen__source_name__source_table_name
+ autogen__source_name__source_table_name__column_name
Note:
- It will use the first path in your config docs-paths or model-paths.
- You can change the docsblock ID prefix (in this case 'autogen'). You can set to None to exclude prefix.
"""
DOCSBLOCK_ID_PREFIX = 'autogen'
DOCSBLOCK_TEMPLATE = '''
{{% docs {docsblock_id} %}}
**\>\>\> INHERITED FROM {source} <<<**
{docsblock_content}
{{% enddocs %}}
'''
def get_dbt_config() -> dict:
with open('dbt_project.yml') as f:
dbt_config = yaml.safe_load(f)
return dbt_config
def get_all_docs_paths(dbt_config: dict) -> list:
all_docs_paths = list(set(
dbt_config.get('docs-paths', [])
+ dbt_config.get('model-paths', [])
)) # use set to deduplicate
return all_docs_paths
def yield_all_yml_files(dbt_config: dict):
all_docs_paths = get_all_docs_paths(dbt_config)
for docs_path in all_docs_paths:
for root, dirs, files in os.walk(docs_path):
for file in files:
if file.endswith(".yml"):
with open(os.path.join(root, file)) as f:
file_data = yaml.safe_load(f)
yield file_data
def render_model_docsblock(model: dict) -> str:
docsblock_id = '__'.join([
DOCSBLOCK_ID_PREFIX,
model.get('name'),
])
source = model.get('name')
docsblock_content = model.get('description', '-') \
.replace('{{', '').replace('}}', '') # avoid jinja inside docs block
return DOCSBLOCK_TEMPLATE.format(
docsblock_id = docsblock_id,
source = source,
docsblock_content = docsblock_content
)
def render_model_column_docsblock(column: dict, model_name: str) -> str:
docsblock_id = '__'.join([
DOCSBLOCK_ID_PREFIX,
model_name,
column.get('name'),
])
source = '.'.join([
model_name,
column.get('name'),
])
docsblock_content = column.get('description', '-') \
.replace('{{', '').replace('}}', '')
return DOCSBLOCK_TEMPLATE.format(
docsblock_id = docsblock_id,
source = source,
docsblock_content = docsblock_content
)
def render_source_table_docsblock(source_table: dict, source_name: str) -> str:
docsblock_id = '__'.join([
DOCSBLOCK_ID_PREFIX,
source_name,
source_table.get('name'),
])
source = '.'.join([
source_name,
source_table.get('name'),
])
docsblock_content = source_table.get('description', '-') \
.replace('{{', '').replace('}}', '')
return DOCSBLOCK_TEMPLATE.format(
docsblock_id = docsblock_id,
source = source,
docsblock_content = docsblock_content
)
def render_source_column_docsblock(column: dict, source_name: str, source_table_name: str) -> str:
docsblock_id = '__'.join([
DOCSBLOCK_ID_PREFIX,
source_name,
source_table_name,
column.get('name'),
])
source = '.'.join([
source_name,
source_table_name,
column.get('name'),
])
docsblock_content = column.get('description', '-') \
.replace('{{', '').replace('}}', '')
return DOCSBLOCK_TEMPLATE.format(
docsblock_id = docsblock_id,
source = source,
docsblock_content = docsblock_content
)
def generate_docsblock():
dbt_config = get_dbt_config()
dbt_docs_path = get_all_docs_paths(dbt_config)[0]
models_docsblock_path = os.path.join(dbt_docs_path, 'autogen_models_docsblock.md')
models_docsblock_file = open(models_docsblock_path, 'w')
sources_docsblock_path = os.path.join(dbt_docs_path, 'autogen_sources_docsblock.md')
sources_docsblock_file = open(sources_docsblock_path, 'w')
for data in yield_all_yml_files(dbt_config):
# render for models
for model in data.get('models', []):
models_docsblock_file.write(render_model_docsblock(model))
for column in model.get('columns', []):
models_docsblock_file.write(render_model_column_docsblock(
column = column,
model_name = model.get('name')
))
# render for sources
for source in data.get('sources', []):
for source_table in source.get('tables', []):
sources_docsblock_file.write(render_source_table_docsblock(
source_table = source_table,
source_name = source.get('name'),
))
for column in source_table.get('columns', []):
sources_docsblock_file.write(render_source_column_docsblock(
column = column,
source_name = source.get('name'),
source_table_name = source_table.get('name')
))
models_docsblock_file.close()
sources_docsblock_file.close()
if __name__ == '__main__':
generate_docsblock()
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment