* On Windows use 127.0.0.1 as default config host (model defaults) and
addionally redirect 0.0.0.0 to localhost on Windows (because default
config file still has 0.0.0.0).
Use 0.0.0.0 as default otherwise (e.g. Linux/Docker) to allow EOS
being accessible on local network (not just same host).
Note: Docs generation on Windows is incompatible with the Github
pipeline tests. Address this in the nested-config feature branch.
* Update install/startup instructions as package installation is
required atm and Docker on Windows has to be accessed at localhost or
127.0.0.1 even though the server log says 0.0.0.0 (which is required
to be available outside the container).
* Fix EOSdash startup with read_only: true (support session key via
EOS_SERVER__EOSDASH_SESSKEY variable). Backport of feature branch.
* Remove root_path, causing Windows to fail load swagger UI (/docs).
* On Windows use 127.0.0.1 as default config host (model defaults) and
addionally redirect 0.0.0.0 to localhost on Windows (because default
config file still has 0.0.0.0).
Use 0.0.0.0 as default otherwise (e.g. Linux/Docker) to allow EOS
being accessible on local network (not just same host).
Note: Docs generation on Windows is incompatible with the Github
pipeline tests. Address this in the nested-config feature branch.
* Update install/startup instructions as package installation is
required atm and Docker on Windows has to be accessed at localhost or
127.0.0.1 even though the server log says 0.0.0.0 (which is required
to be available outside the container).
* Fix EOSdash startup with read_only: true (support session key via
EOS_SERVER__EOSDASH_SESSKEY variable). Backport of feature branch.
* Remove root_path, causing Windows to fail load swagger UI (/docs).
* /v1/config/{path} supports setting single config value (post body). Lists are
supported as well by using the index:
- general/latitude (value: 55.55)
- optimize/ev_available_charge_rates_percent/0 (value: 42)
Whole tree can be overriden as well (no merge):
- optimize/ev_available_charge_rates_percent (value: [42, 43, 44]
* ConfigEOS: Add set_config_value, get_config_value
* Docker: Copy the same directory structure (src/) to support the
lifespan startup of EOSdash.
Use EOS_SERVER_EOSDASH_SESSKEY environment variable to provide
EOSdash with session key.
* All config now nested.
- Use default config from model field default values. If providers
should be enabled by default, non-empty default config file could
be provided again.
- Environment variable support with EOS_ prefix and __ between levels,
e.g. EOS_SERVER__EOS_SERVER_PORT=8503 where all values are case
insensitive.
For more information see:
https://docs.pydantic.dev/latest/concepts/pydantic_settings/#parsing-environment-variable-values
- Use devices as registry for configured devices. DeviceBase as base
class with for now just initializion support (in the future expand
to operations during optimization).
- Strip down ConfigEOS to the only configuration instance. Reload
from file or reset to defaults is possible.
* Fix multi-initialization of derived SingletonMixin classes.
* Cleanup: parameters: extra=forbid, optimize: battery, inverter optional
* Don't allow extra fields for parameters/REST-API (at least for now while
changing API).
* Allow both battery and inverter to be set optionally (atm optional
battery not implemented, no API constraints).
* inverter: Remove default max_power_wh
* single_test_optimization: Add more cli-parameters
* Workflow docker-build: Don't try to authenticate for PRs
* Secrets are not available anyway for forks.
* workflow: docker-build upload to DockerHub
- Upload on release, tag, push to main.
- Build on pr to main (amd64 only).
* docker:
- Update documentation.
- Temporarily set akkudoktor/eos:main in compose.yml (with
releases/tags it should be replaced by latest again)
* workflow: docker-build upload to DockerHub
- Upload on release, tag, push to main.
- Build on pr to main (amd64 only).
* docker: temporarily disable read-only container Closes#278
- Update documentation.
- Temporarily set akkudoktor/eos:main in compose.yml (with
releases/tags it should be replaced by latest again)
* Provide /v1/prediction/update and
/v1/prediction/update/{provider_id} as POST endpoints to update
provider data.
* Update description for deprecated endpoints how to use new API.
* Add EOS_CONFIG_DIR to set config dir (relative path to EOS_DIR or
absolute path).
- config_folder_path read-only
- config_file_path read-only
* Default values to support app start with empty config:
- latitude/longitude (Berlin)
- optimization_ev_available_charge_rates_percent (null, so model
default value is used)
- Enable Akkudoktor electricity price forecast (docker-compose).
* Fix some endpoints (empty data, remove unused params, fix types).
* cacheutil: Use cache dir. Closes#240
* Support EOS_LOGGING_LEVEL environment variable to set log level.
* tests: All tests use separate temporary config
- Add pytest switch --check-config-side-effect to check user
config file existence after each test. Will also fail if user config
existed before test execution (but will only check after the test has
run).
Enable flag in github workflow.
- Globally mock platformdirs in config module. Now no longer required
to patch individually.
Function calls to config instance (e.g. merge_settings_from_dict)
were unaffected previously.
* Set Berlin as default location (default config/docker-compose).
* optimization states for AC, DC and IDLE now similar probab. Also AC states taken from config. Maybe a single config option for AC and E-Auto States is sensefull.
* Ruff
* test_class_optimize: Update testdata
* Write pdf and json to test/testdata/new.... so it can be analyzed
manually or just copied as new expected result.
* workflow: Upload pytest optimization result artifacts (pdf, json)
---------
Co-authored-by: Andreas <drbacke@gmx.de>
* Install missing deps. In the future we probably just want to reuse
the dev environment (e.g. poetry shell) to have same result for hook
and explicit execution in venv.
* Mypy: Initial support
* Add to pre-commit (currently installs own deps, could maybe changed
to poetry venv in the future to reuse environment and don't need
duplicated types deps).
* Add type hints.
* Mypy: Add missing annotations
* Update README/CONTRIBUTING Closes#180
* Remove duplicated files in docs/develop and copy them from project
root to there on documentation generation.
* Makefile: Use fastapi to start server, set port to 8503
* Migrate from Flask to FastAPI
* FastAPI migration:
- Use pydantic model classes as input parameters to the
data/calculation classes.
- Interface field names changed to constructor parameter names (for
simplicity only during transition, should be updated in a followup
PR).
- Add basic interface requirements (e.g. some values > 0, etc.).
* Update tests for new data format.
* Python requirement down to 3.9 (TypeGuard no longer needed)
* Makefile: Add helpful targets (e.g. development server with reload)
* Move API doc from README to pydantic model classes (swagger)
* Link to swagger.io with own openapi.yml.
* Commit openapi.json and check with pytest for changes so the
documentation is always up-to-date.
* Streamline docker
* FastAPI: Run startup action on dev server
* Fix config for /strompreis, endpoint still broken however.
* test_openapi: Compare against docs/.../openapi.json
* Move fastapi to server/ submodule
* See #187 for new repository structure.
* Integrated single_test_optimization into pytest to run a basic optimization test with tolerance set to 1e-6, ensuring quick detection of deviations.
* Added a long-run test (400 generations, like single_test_optimization), which can be triggered using --full-run in pytest.
* Mocked PDF creation in optimization tests and added a new PDF generation test with image comparison validation.
Note: Current tolerance is set to 1e-6; feedback on whether this tolerance is tight enough is welcome.
---------
Co-authored-by: Normann <github@koldrack.com>
Co-authored-by: Michael Osthege <michael.osthege@outlook.com>
* Dockerfile: Use non-root user, buildx cache, setup for readonly
container, remove unused apt deps.
For now don't install pip package and keep development flask server
as this will be replaced in the future (fastapi). Then a proper
webserver (e.g. nginx) should be used and the pip package can be
created and deployed just to the run-stage (with the webserver).
* docker-compose: Set to readonly (anonymous volumes declared in
Dockerfile should maintain all writable data).
Mount config.py for easier development. Should be replaced by
environment support for all config file variables.
* Remove unused runtime dependencies: mariadb, joblib, pytest,
pytest-cov.
* Move pytest-cov to dev dependencies.
* Add output_dir to config.py.
* Fix visualization_results.pdf endpoint.
* Update docs.
