🚧 dlt-init-openapi demo

Generates dlt pipelines from OpenAPI 3.x documents.

This generator does not support OpenAPI 2.x FKA Swagger. If you need to use an older document, try upgrading it to version 3 first with one of many available converters.

🕳️ This is only a demo.

will generate resources for all endpoints that return lists of objects

will use a few heuristics to find list wrapped in responses

will generate transformers from all endpoints that have a matching list resource (same object type returned)

will use a few heuristics to find the right object id to pass to the transformer

user can select endpoints using questionary lib in CLI

endpoints that have the most central data types (tables linking to many other tables) will be listed first

the structure of the code is not optimized!

there's no pagination added. use our GPT-4 playground to do that

Generating a Pokemon dlt pipeline from Open API Spec 🚀

Prior work

This is a heavily hacked fork of openapi-python-client

Usage

You need poetry to install dependencies

poetry install
poetry shell

Create new dlt pipeline from PokeAPI spec and place it in the pokemon-pipeline

dlt-init init pokemon --url https://raw.githubusercontent.com/cliffano/pokeapi-clients/main/specification/pokeapi.yml

After executing of the command, you can pick the endpoints that you want to add to your source and then load with the pipeline. The endpoints are grouped by returned data type (table) and ordered by centrality (a measure how many other tables, the given table links to):

? Which resources would you like to generate? (Use arrow keys to move, <space> to select, <a> to toggle, <i> to invert)
 
PokemonSpecies endpoints:

   ○ pokemon_species_list /api/v2/pokemon-species/
 » ○ pokemon_species_read /api/v2/pokemon-species/{id}/
 
EvolutionChain endpoints:

   ○ evolution_chain_list /api/v2/evolution-chain/
   ○ evolution_chain_read /api/v2/evolution-chain/{id}/
 
MoveAilment endpoints:

   ○ move_ailment_list /api/v2/move-ailment/
   ○ move_ailment_read /api/v2/move-ailment/{id}/
 
Move endpoints:

   ○ move_list /api/v2/move/
   ○ move_read /api/v2/move/{id}/
 
Pokemon endpoints:

   ○ pokemon_list /api/v2/pokemon/
   ○ pokemon_read /api/v2/pokemon/{id}/

Pick your endpoints and press ENTER to generate pipeline. Now you are ready to load data.
Enter the pokemon-pipeline folder and execute the pipeline.py script. This will load your endpoints to local duckdb. Below we use enlighten to show fancy progress bars:

cd pokemon-pipeline
PROGRESS=enlighten python pipeline.py

Inspect the pipeline to see what got loaded

$ dlt pipeline pokemon_pipeline info
Found pipeline pokemon_pipeline in /home/rudolfix/.dlt/pipelines
Synchronized state:
_state_version: 2
_state_engine_version: 2
pipeline_name: pokemon_pipeline
dataset_name: pokemon_data
default_schema_name: pokemon
schema_names: ['pokemon']
destination: dlt.destinations.duckdb

Local state:
first_run: False
_last_extracted_at: 2023-06-12T11:50:16.171872+00:00

Resources in schema: pokemon
pokemon_species_read with 8 table(s) and 0 resource state slot(s)

Working dir content:
Has 1 completed load packages with following load ids:
1686570616.17882

Pipeline has last run trace. Use 'dlt pipeline pokemon_pipeline trace' to inspect

Launch the streamlit app to preview the data (we copy a streamlit config to make it work on codespaces)

cp -r ../.streamlit .
pip install pandas streamlit
dlt pipeline pokemon_pipeline show

What You Get

When you run the command above, following files will be generated:

pokemon-pipeline a folder with all the files
a folder pokemon with the Python module containing dlt source, resources and the Python client.
__init__.py in that folder with the dlt source
the pipeline.py file that loads the resources to duckdb
.dlt folder with the config.toml

What's next

There's still work needed to make things useful:

We will fully restructure the underlying Python client. We'll compress all the files in pokemon/api folder into a single, nice and extendable client.
We'll allow to easily add pagination and other injections into client. GPT-4 friendly
Many more heuristics to extract resources and their dependencies
Integration with existing dlt init command

🚀 openapi-python-client docs

If you want to experiment, features below still work

OpenAPI features supported

All HTTP Methods
JSON and form bodies, path and query parameters
File uploads with multipart/form-data bodies
float, string, int, date, datetime, string enums, and custom schemas or lists containing any of those
html/text or application/json responses containing any of the previous types
Bearer token security

Configuration

You can pass a YAML (or JSON) file to openapi-python-client with the --config option in order to change some behavior. The following parameters are supported:

project_name_override and package_name_override

Pass the source in command line to create pipeline instead!

field_prefix

When generating properties, the name attribute of the OpenAPI schema will be used. When the name is not a valid Python identifier (e.g. begins with a number) this string will be prepended. Defaults to "field_". It will also be used to prefix fields in schema starting with "_" in order to avoid ambiguous semantics.

Example:

field_prefix: attr_

package_version_override

Specify the package version of the generated client. If unset, the client will use the version of the OpenAPI spec.

Example:

package_version_override: 1.2.3

post_hooks

In the config file, there's an easy way to tell openapi-python-client to run additional commands after generation. Here's an example showing the default commands that will run if you don't override them in config:

post_hooks:
   - "autoflake -i -r --remove-all-unused-imports --remove-unused-variables --ignore-init-module-imports ."
   - "isort ."
   - "black ."

http_timeout

By default, the timeout for retrieving the schema file via HTTP is 5 seconds. In case there is an error when retrieving the schema, you might try and increase this setting to a higher value.

dlt-hub / dlt-init-openapi