Qxmt

ADD

FIX
- Process parallelization process. Mainly, when a file is changed, the latest code is used in the child process.

CHANGE

Add
- record the runtime of fit and predict in each RunRecord (`"runtime": {"fit": 3.2, "predict": 1.2}`).

Fix
- Kernel build method from feature map function.
- Disable cache when reloading custom modules.
- Temporarily changed to single-threaded due to a bug found in the process parallelization process. Will be supported in the future.

Change
- remove input dim check when construct feature map. Implement user's own necessary checking mechanisms to improve operability.
- split default feature map module (`defaults.py`) in to each objective and method.

ADD
- check git availability (106 )
- add file loader (108)
- generated data loader (109)
- Infer dataset type (111 )
- support for sequential processing of multi logic (112)
- sampling mode for fidelity kernel (114)
- save result summary of shots (115)

FIX
- unify graph color (113)

Change
- rename dummy module to generate module (109)
- remove params from dataset config (110)

Add
- hyperparamter search by optuna (95)
- support multi class (96)
- quantum rigge regressor class (97)
- split evaluation class into clf or reg (99)
- visualization method for regression model (100)
- generate method for dummy regression data (101)
- cross_val and hyperparams-search for QRigge
- implement QSVR model (103)
- support py310 environment (104)

Fix
- changed user environment path when executing custom modules (93)
- rename QSVM to QSVC (102)

Add
* (add): split logic of validation (https://github.com/Qyusu/qxmt/pull/80)

* (add): cross validation (https://github.com/Qyusu/qxmt/pull/81)

* (add): role for description generator (https://github.com/Qyusu/qxmt/pull/83)

* (add): dataset load from OpenML (https://github.com/Qyusu/qxmt/pull/84)

* (add): config for multi dataset type (https://github.com/Qyusu/qxmt/pull/85)

* (add): override empty experiment directory (https://github.com/Qyusu/qxmt/pull/85)

* (add): visualize circuit in the MPL format (https://github.com/Qyusu/qxmt/pull/86)

* (add): projected kernel (https://github.com/Qyusu/qxmt/pull/86)

* (add): hyper-parameter search method (grid, random) (https://github.com/Qyusu/qxmt/pull/90)

Change
* (change): move raw dataset type definition (https://github.com/Qyusu/qxmt/pull/80)

* (change): definition of default metrics in one place (https://github.com/Qyusu/qxmt/pull/82)

Fix
* (fix): dummy dataset label (https://github.com/Qyusu/qxmt/pull/82)

Docs
* Change structure of API reference page (https://github.com/Qyusu/qxmt/pull/91)
* Support tutorial in v0.2.3 (https://github.com/Qyusu/qxmt/pull/91)
* Add contribution guideline (https://github.com/Qyusu/qxmt/pull/82)

What's Changed
* V0.2.1 dev by kenya-sk in https://github.com/Qyusu/qxmt/pull/70
* (docs): deploy by actions by kenya-sk in https://github.com/Qyusu/qxmt/pull/74
* (docs): add custom method tutorial by kenya-sk in https://github.com/Qyusu/qxmt/pull/75
* (docs): update workflow by kenya-sk in https://github.com/Qyusu/qxmt/pull/76
* (add): predict_proba method for qsvm by kenya-sk in https://github.com/Qyusu/qxmt/pull/77


**Full Changelog**: https://github.com/Qyusu/qxmt/compare/v0.2.1...v0.2.2