Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[D] Stable baselines integration #384

Merged
merged 38 commits into from
Apr 8, 2022
Merged

Conversation

jannikgro
Copy link
Collaborator

@jannikgro jannikgro commented Apr 4, 2022

This PR aims to integrate stable baselines (closes #316). All needed functionality is delivered, but there is an ugly code duplication between stable_baselines_callback and training. It is an option (and my preferred one) to refactor this code in a way that the callback is called in our agents too. But this solution needs some bit efford and could be written as a new issue.

@jannikgro jannikgro added the research A feature offering value to Alex label Apr 4, 2022
@jannikgro jannikgro self-assigned this Apr 4, 2022
@jannikgro jannikgro changed the title Stable baselines integration [D] Stable baselines integration Apr 4, 2022
@jannikgro jannikgro added the discuss To be discussed label Apr 5, 2022
@jannikgro jannikgro requested review from NikkelM and ldreessen April 5, 2022 14:56
@jannikgro jannikgro marked this pull request as ready for review April 5, 2022 14:56
Copy link
Collaborator

@NikkelM NikkelM left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You wrote that you want to do a refactoring to remove some code duplication. To keep in line with our motto of keeping PR's small, I would prefer if you did so in a separate PR, perhaps right after this one is merged.


In one of my comments I mentioned a bug regarding the _on_training_end, I am currently user-testing if this also occurs for our other trainings or is limited to the StableBaselines-Training.
Update: Yes, the bug I mentioned also occurred when I trained a QLearningAgent
Additional edit: It seems like this was due to an unset datapath, so we should perhaps include a check that the datapath is set before starting the training

@jannikgro jannikgro requested review from blackjack2693 and removed request for ldreessen April 7, 2022 15:19
@jannikgro jannikgro merged commit f8bc162 into development Apr 8, 2022
@jannikgro jannikgro deleted the stable_baselines_integration branch April 8, 2022 11:20
felix-20 added a commit that referenced this pull request Apr 11, 2022
commit 091eee6
Author: Nikkel Mollenhauer <[email protected]>
Date:   Mon Apr 11 10:43:00 2022 +0200

    Tests for `config_validation.py` (#404)

    * Don't get default data (shouldn't be necessary)

    * Unpack default data

    * Refactored utils functions to return dicts instead of strings

    * Adapted to new mock format

    * Adapted to new mock format

    * Some first tests

    * More tests

    * Fixed testcase-names

    * Moved file-endings to initial function call

    * Fixed tests

    * More asserts

    * More tests

    * More tests

    * More tests

    * `validate_sub_keys`-tests

commit f8bc162
Author: jannikgro <[email protected]>
Date:   Fri Apr 8 13:20:42 2022 +0200

    [D] Stable baselines integration (#384)

    * refactored reinforcement learning agent to accept marketplace

    * adapted test_exampleprinter.py to marketplace initialization

    * add market option to accept continuos actions

    * fixed action space check

    * initial stable baselines integration

    * Agent init by env (#390)

    * introduced self.network in actorcritic_agent

    * added network_architecture in QLearningAgent

    * changed actorcritic_agent to network_architecture

    * set back training_scenario

    * am_configuration initialize rl-agent via marketplace

    * added final analyse to stable baselines training

    * added more stable baselines algorithms

    * added ppo algorithm

    * introduced stable_baselines_folder

    * renamed training to callback

    * satisfied linter

    * fixed loading problem

    * try to make tqdm run in stable_baselines

    * make tqdm running

    * reduced pmonitoring episodes in sb training

    * save model only if significantly better

    * fixed too long test time bug

    * moved back to 250 episodes testing

    * set timeout to 15 minutes

    * added first batch of fixes to @NikkelM feedback

    * added type annotations and asserts in stable_baselines_model

    * added sbtraining to training_scenario

    * applied comments in am_configuration

    * solved .dat problem and fixed crashing asserts

    * reintroduced _end_of_training

    * removed deprecated if

    * Moved '.dat' to function call instead of appending within function

    * Fixed assert

    * fixed model file ending bug

    * Add short explanation docstring

    Co-authored-by: Johann Schulze Tast <[email protected]>

    * fixed wrong docstring

    * Fixed tests

    Co-authored-by: NikkelM <[email protected]>
    Co-authored-by: Johann Schulze Tast <[email protected]>
felix-20 added a commit that referenced this pull request Apr 12, 2022
* ability to add `DockerInfo` for multiple container

* support for starting multiple container on docker side

* multiple experiments are supported on the webserver

* fix #380

* name from `names` for container

* merge development

commit 091eee6
Author: Nikkel Mollenhauer <[email protected]>
Date:   Mon Apr 11 10:43:00 2022 +0200

    Tests for `config_validation.py` (#404)

    * Don't get default data (shouldn't be necessary)

    * Unpack default data

    * Refactored utils functions to return dicts instead of strings

    * Adapted to new mock format

    * Adapted to new mock format

    * Some first tests

    * More tests

    * Fixed testcase-names

    * Moved file-endings to initial function call

    * Fixed tests

    * More asserts

    * More tests

    * More tests

    * More tests

    * `validate_sub_keys`-tests

commit f8bc162
Author: jannikgro <[email protected]>
Date:   Fri Apr 8 13:20:42 2022 +0200

    [D] Stable baselines integration (#384)

    * refactored reinforcement learning agent to accept marketplace

    * adapted test_exampleprinter.py to marketplace initialization

    * add market option to accept continuos actions

    * fixed action space check

    * initial stable baselines integration

    * Agent init by env (#390)

    * introduced self.network in actorcritic_agent

    * added network_architecture in QLearningAgent

    * changed actorcritic_agent to network_architecture

    * set back training_scenario

    * am_configuration initialize rl-agent via marketplace

    * added final analyse to stable baselines training

    * added more stable baselines algorithms

    * added ppo algorithm

    * introduced stable_baselines_folder

    * renamed training to callback

    * satisfied linter

    * fixed loading problem

    * try to make tqdm run in stable_baselines

    * make tqdm running

    * reduced pmonitoring episodes in sb training

    * save model only if significantly better

    * fixed too long test time bug

    * moved back to 250 episodes testing

    * set timeout to 15 minutes

    * added first batch of fixes to @NikkelM feedback

    * added type annotations and asserts in stable_baselines_model

    * added sbtraining to training_scenario

    * applied comments in am_configuration

    * solved .dat problem and fixed crashing asserts

    * reintroduced _end_of_training

    * removed deprecated if

    * Moved '.dat' to function call instead of appending within function

    * Fixed assert

    * fixed model file ending bug

    * Add short explanation docstring

    Co-authored-by: Johann Schulze Tast <[email protected]>

    * fixed wrong docstring

    * Fixed tests

    Co-authored-by: NikkelM <[email protected]>
    Co-authored-by: Johann Schulze Tast <[email protected]>

* clean up docker port

* fix docker test

* docstrings

* implement #356

* missed `handle_requests` on last commit

* Merge branch 'development' into 379-multiple-experiments
felix-20 added a commit that referenced this pull request Apr 20, 2022
commit 8652a67
Author: Judith <[email protected]>
Date:   Mon Apr 11 18:17:20 2022 +0200

    missed `handle_requests` on last commit

commit a354f89
Author: Judith <[email protected]>
Date:   Mon Apr 11 18:16:15 2022 +0200

    implement #356

commit fd54365
Author: Judith <[email protected]>
Date:   Mon Apr 11 17:56:35 2022 +0200

    docstrings

commit 9f43c9d
Author: Judith <[email protected]>
Date:   Mon Apr 11 15:07:47 2022 +0200

    fix docker test

commit b3e6e80
Merge: 6f64f9e 091eee6
Author: felix-20 <[email protected]>
Date:   Mon Apr 11 14:31:39 2022 +0200

    Merge branch 'development' into 379-multiple-experiments

commit 6f64f9e
Author: Judith <[email protected]>
Date:   Mon Apr 11 14:20:47 2022 +0200

    clean up docker port

commit 8c35e8a
Author: Judith <[email protected]>
Date:   Mon Apr 11 11:33:30 2022 +0200

    merge development

    commit 091eee6
    Author: Nikkel Mollenhauer <[email protected]>
    Date:   Mon Apr 11 10:43:00 2022 +0200

        Tests for `config_validation.py` (#404)

        * Don't get default data (shouldn't be necessary)

        * Unpack default data

        * Refactored utils functions to return dicts instead of strings

        * Adapted to new mock format

        * Adapted to new mock format

        * Some first tests

        * More tests

        * Fixed testcase-names

        * Moved file-endings to initial function call

        * Fixed tests

        * More asserts

        * More tests

        * More tests

        * More tests

        * `validate_sub_keys`-tests

    commit f8bc162
    Author: jannikgro <[email protected]>
    Date:   Fri Apr 8 13:20:42 2022 +0200

        [D] Stable baselines integration (#384)

        * refactored reinforcement learning agent to accept marketplace

        * adapted test_exampleprinter.py to marketplace initialization

        * add market option to accept continuos actions

        * fixed action space check

        * initial stable baselines integration

        * Agent init by env (#390)

        * introduced self.network in actorcritic_agent

        * added network_architecture in QLearningAgent

        * changed actorcritic_agent to network_architecture

        * set back training_scenario

        * am_configuration initialize rl-agent via marketplace

        * added final analyse to stable baselines training

        * added more stable baselines algorithms

        * added ppo algorithm

        * introduced stable_baselines_folder

        * renamed training to callback

        * satisfied linter

        * fixed loading problem

        * try to make tqdm run in stable_baselines

        * make tqdm running

        * reduced pmonitoring episodes in sb training

        * save model only if significantly better

        * fixed too long test time bug

        * moved back to 250 episodes testing

        * set timeout to 15 minutes

        * added first batch of fixes to @NikkelM feedback

        * added type annotations and asserts in stable_baselines_model

        * added sbtraining to training_scenario

        * applied comments in am_configuration

        * solved .dat problem and fixed crashing asserts

        * reintroduced _end_of_training

        * removed deprecated if

        * Moved '.dat' to function call instead of appending within function

        * Fixed assert

        * fixed model file ending bug

        * Add short explanation docstring

        Co-authored-by: Johann Schulze Tast <[email protected]>

        * fixed wrong docstring

        * Fixed tests

        Co-authored-by: NikkelM <[email protected]>
        Co-authored-by: Johann Schulze Tast <[email protected]>

commit 0183876
Author: Judith <[email protected]>
Date:   Mon Apr 11 11:31:15 2022 +0200

    name from `names` for container

commit ad70b7c
Author: Judith <[email protected]>
Date:   Sun Apr 10 16:32:47 2022 +0200

    fix #380

commit 8e1314c
Author: Judith <[email protected]>
Date:   Sun Apr 10 16:05:28 2022 +0200

    multiple experiments are supported on the webserver

commit 53a4000
Author: Judith <[email protected]>
Date:   Fri Apr 8 14:36:44 2022 +0200

    support for starting multiple container on docker side

commit 3c94f43
Author: Judith <[email protected]>
Date:   Fri Apr 8 11:46:07 2022 +0200

    ability to add `DockerInfo` for multiple container
felix-20 added a commit that referenced this pull request Jul 20, 2022
* merge multiple experiments

commit 8652a67
Author: Judith <[email protected]>
Date:   Mon Apr 11 18:17:20 2022 +0200

    missed `handle_requests` on last commit

commit a354f89
Author: Judith <[email protected]>
Date:   Mon Apr 11 18:16:15 2022 +0200

    implement #356

commit fd54365
Author: Judith <[email protected]>
Date:   Mon Apr 11 17:56:35 2022 +0200

    docstrings

commit 9f43c9d
Author: Judith <[email protected]>
Date:   Mon Apr 11 15:07:47 2022 +0200

    fix docker test

commit b3e6e80
Merge: 6f64f9e 091eee6
Author: felix-20 <[email protected]>
Date:   Mon Apr 11 14:31:39 2022 +0200

    Merge branch 'development' into 379-multiple-experiments

commit 6f64f9e
Author: Judith <[email protected]>
Date:   Mon Apr 11 14:20:47 2022 +0200

    clean up docker port

commit 8c35e8a
Author: Judith <[email protected]>
Date:   Mon Apr 11 11:33:30 2022 +0200

    merge development

    commit 091eee6
    Author: Nikkel Mollenhauer <[email protected]>
    Date:   Mon Apr 11 10:43:00 2022 +0200

        Tests for `config_validation.py` (#404)

        * Don't get default data (shouldn't be necessary)

        * Unpack default data

        * Refactored utils functions to return dicts instead of strings

        * Adapted to new mock format

        * Adapted to new mock format

        * Some first tests

        * More tests

        * Fixed testcase-names

        * Moved file-endings to initial function call

        * Fixed tests

        * More asserts

        * More tests

        * More tests

        * More tests

        * `validate_sub_keys`-tests

    commit f8bc162
    Author: jannikgro <[email protected]>
    Date:   Fri Apr 8 13:20:42 2022 +0200

        [D] Stable baselines integration (#384)

        * refactored reinforcement learning agent to accept marketplace

        * adapted test_exampleprinter.py to marketplace initialization

        * add market option to accept continuos actions

        * fixed action space check

        * initial stable baselines integration

        * Agent init by env (#390)

        * introduced self.network in actorcritic_agent

        * added network_architecture in QLearningAgent

        * changed actorcritic_agent to network_architecture

        * set back training_scenario

        * am_configuration initialize rl-agent via marketplace

        * added final analyse to stable baselines training

        * added more stable baselines algorithms

        * added ppo algorithm

        * introduced stable_baselines_folder

        * renamed training to callback

        * satisfied linter

        * fixed loading problem

        * try to make tqdm run in stable_baselines

        * make tqdm running

        * reduced pmonitoring episodes in sb training

        * save model only if significantly better

        * fixed too long test time bug

        * moved back to 250 episodes testing

        * set timeout to 15 minutes

        * added first batch of fixes to @NikkelM feedback

        * added type annotations and asserts in stable_baselines_model

        * added sbtraining to training_scenario

        * applied comments in am_configuration

        * solved .dat problem and fixed crashing asserts

        * reintroduced _end_of_training

        * removed deprecated if

        * Moved '.dat' to function call instead of appending within function

        * Fixed assert

        * fixed model file ending bug

        * Add short explanation docstring

        Co-authored-by: Johann Schulze Tast <[email protected]>

        * fixed wrong docstring

        * Fixed tests

        Co-authored-by: NikkelM <[email protected]>
        Co-authored-by: Johann Schulze Tast <[email protected]>

commit 0183876
Author: Judith <[email protected]>
Date:   Mon Apr 11 11:31:15 2022 +0200

    name from `names` for container

commit ad70b7c
Author: Judith <[email protected]>
Date:   Sun Apr 10 16:32:47 2022 +0200

    fix #380

commit 8e1314c
Author: Judith <[email protected]>
Date:   Sun Apr 10 16:05:28 2022 +0200

    multiple experiments are supported on the webserver

commit 53a4000
Author: Judith <[email protected]>
Date:   Fri Apr 8 14:36:44 2022 +0200

    support for starting multiple container on docker side

commit 3c94f43
Author: Judith <[email protected]>
Date:   Fri Apr 8 11:46:07 2022 +0200

    ability to add `DockerInfo` for multiple container

* first attempt for websocket

* websocket on docker site working

* just send all of it if things have changed

* started on webserver site

* webserver sends push notification to user about stopped container

* commit

* websocket to ssl

* fix

* try

* started on database manager

* more db

* fix?

* debug statements

* more types

* does table exist fix

* colorful terminal output

* commit

* debug

* extra health checker

* debug for force stop

* some logging

* fix tests

* telegram notifications

* better logging

* better logging

* fix test?

* different logging level

* small fixes

* started on system monitoring

* system monitoring

* csv possibility for system data

* two buttons in webserver

* more fail prove

* fix configuration form

* silent_starter

* logging statements for silent starter

* debug

* debug

* more debugging

* fix error

* more debugging

* more local monitoring

* adjusted logging in silent starter

* gpu

* os.system

* subprocess

* Squashed commit of the following:

commit af12e13
Merge: bc9d3a3 ab704ca
Author: Nikkel Mollenhauer <[email protected]>
Date:   Tue Jun 7 18:20:56 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit bc9d3a3
Author: Nikkel Mollenhauer <[email protected]>
Date:   Tue Jun 7 18:20:50 2022 +0200

    Removed dead methods

commit ab704ca
Author: felix-20 <[email protected]>
Date:   Tue Jun 7 16:55:34 2022 +0200

    improved prefill (#496)

    * improved prefill with consideration of the current formdata

    * remove debug statements

    * remove print debug

    * Added lxml dependency

    * Removed debug comment

    Co-authored-by: Nikkel Mollenhauer <[email protected]>

commit 8839f90
Author: Judith <[email protected]>
Date:   Tue Jun 7 12:01:43 2022 +0200

    remove `config_is_final`

commit 53d79f9
Author: NikkelM <[email protected]>
Date:   Tue Jun 7 10:41:51 2022 +0200

    Review feedback by @SinNeax

commit 99ab681
Author: NikkelM <[email protected]>
Date:   Sat Jun 4 15:02:34 2022 +0200

    Fixed config validation

commit 6a09267
Merge: 179e0da e623413
Author: Judith <[email protected]>
Date:   Fri Jun 3 14:50:15 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit 179e0da
Author: Judith <[email protected]>
Date:   Fri Jun 3 14:49:49 2022 +0200

    fix javascript error

commit e623413
Author: NikkelM <[email protected]>
Date:   Fri Jun 3 14:38:20 2022 +0200

    Added `config_type` field to default modelfiles

commit bd15108
Merge: dac92d3 507cbc1
Author: NikkelM <[email protected]>
Date:   Fri Jun 3 09:45:11 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit dac92d3
Author: NikkelM <[email protected]>
Date:   Fri Jun 3 09:44:49 2022 +0200

    Fixed tests

commit 507cbc1
Author: Judith <[email protected]>
Date:   Fri Jun 3 07:52:30 2022 +0200

    dynamic table for sim_market possible

commit bce03bd
Author: Judith <[email protected]>
Date:   Thu Jun 2 20:15:17 2022 +0200

    fix webserver tests

commit 57dbb33
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 14:29:40 2022 +0200

    Webserver format

commit b3780cd
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 14:15:01 2022 +0200

    Fixed validation

commit c9657a9
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 13:50:20 2022 +0200

    Added back needed functionality of validating "complete" configs

commit dca24f3
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 13:34:21 2022 +0200

    Added some tests

commit fff9b47
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 13:12:41 2022 +0200

    Renamed `market` to `sim_market`

commit caf5c01
Merge: b8c43b7 f50fe65
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 11:30:18 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit b8c43b7
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 11:29:34 2022 +0200

    New config format

commit f50fe65
Author: Judith <[email protected]>
Date:   Thu Jun 2 11:21:10 2022 +0200

    implement new config validation for webserver

commit 135e8f3
Author: Judith <[email protected]>
Date:   Wed Jun 1 20:52:12 2022 +0200

    fix webserver tests

commit d326245
Author: Judith <[email protected]>
Date:   Wed Jun 1 17:26:50 2022 +0200

    fix prefill

commit 70492c8
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 17:08:34 2022 +0200

    Fixed remaining tests(?)

commit 59136b1
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 16:35:00 2022 +0200

    Fixed most tests

commit c912309
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 16:11:20 2022 +0200

    Removed references to `class`-field

commit 2ccee78
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 15:12:56 2022 +0200

    Removed `class` keywords from config files

commit c51dab7
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 15:09:46 2022 +0200

    renamed rl_config to q_learning_config

commit f9aa012
Author: NikkelM <[email protected]>
Date:   Tue May 31 19:49:14 2022 +0200

    Fixed Agent_monitoring with stable_baselines agents

commit 16df8c4
Merge: 7610c6e 26cb54b
Author: Judith <[email protected]>
Date:   Tue May 31 13:50:24 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit 7610c6e
Author: Judith <[email protected]>
Date:   Tue May 31 13:50:14 2022 +0200

    script for creating rl model

commit 26cb54b
Author: NikkelM <[email protected]>
Date:   Tue May 31 11:21:46 2022 +0200

    Simplified and extended key validation

commit 0fa7eb5
Author: Judith <[email protected]>
Date:   Mon May 30 14:17:22 2022 +0200

    rl config works in view

commit da05474
Author: NikkelM <[email protected]>
Date:   Fri May 27 13:37:39 2022 +0200

    New feather website

commit dd40104
Author: NikkelM <[email protected]>
Date:   Fri May 27 12:56:28 2022 +0200

    Removed debug comment, fixed policyanalyzer

commit 1d025c9
Author: NikkelM <[email protected]>
Date:   Fri May 27 12:39:17 2022 +0200

    Reintroduced `main` for testing purposes

commit c90a8d6
Author: NikkelM <[email protected]>
Date:   Fri May 27 12:26:10 2022 +0200

    Fixed config validation for webserver

commit 4107aaf
Author: NikkelM <[email protected]>
Date:   Fri May 27 10:57:04 2022 +0200

    Added some debugging to docker

commit 2588f6d
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 22:22:39 2022 +0200

    preconditions for actor critic parameters and stable baselines parameter have been created

commit 4e174e6
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:42:23 2022 +0200

    restored changes

commit 14f181d
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:37:23 2022 +0200

    restored last changes

commit 3d3ac9d
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:34:11 2022 +0200

    more information

commit 31a212f
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:31:31 2022 +0200

    more information

commit 1b139e0
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:28:02 2022 +0200

    test

commit 1e80e8c
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:25:16 2022 +0200

    next try

commit 09a433e
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:22:45 2022 +0200

    another printf try

commit de9f5eb
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:19:07 2022 +0200

    more printf debug

commit 511d38a
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:16:11 2022 +0200

    printf debug

commit 1d1a3c4
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:12:06 2022 +0200

    restored last changes

commit 61c0d8e
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:09:22 2022 +0200

    tried without reading check

commit 068abaa
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:03:54 2022 +0200

    tried with dirty fix

commit f1cdf45
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:02:03 2022 +0200

    debug print

commit 75a674e
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 18:56:08 2022 +0200

    restored docker_manager

commit 39bc238
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 18:51:31 2022 +0200

    tried with assert False

commit 6986513
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 18:21:59 2022 +0200

    small change in docker_manager

commit 561fc00
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 18:13:18 2022 +0200

    tried to fix config_validation

commit 3fded29
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 17:25:10 2022 +0200

    reintroduced test_hyperparameter_config_market

commit b6a4945
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 15:52:19 2022 +0200

    reintroduced test_hyperparameter_config_rl

commit 7cef3e7
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 15:19:29 2022 +0200

    added rules and verifications

commit 473ddf7
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 12:43:12 2022 +0200

    introduced JSONConfigurable class and demanded class entry in config

commit 36024c8
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 11:50:48 2022 +0200

    solved #484 without check

* automatically write template files

* try fix pre-commit

* try fix pre-commit

* try except

* more debug

* more debug

* debugging :)

* hotfix for qlearning file

* adopting to the given agent works again

* restore docker manager and app.py

* try restore again

* webserver tests are running again

merge from dev was very very strange

* new rl_config

* precommit?

* precommit!

* assert print

Co-Authored-By: Nikkel Mollenhauer <[email protected]>

* remove print

Co-Authored-By: Nikkel Mollenhauer <[email protected]>

* new regex for matching ce agents

Co-Authored-By: Nikkel Mollenhauer <[email protected]>

* fixed fixed price agent

Co-Authored-By: Nikkel Mollenhauer <[email protected]>

* comment out assert

Co-Authored-By: Nikkel Mollenhauer <[email protected]>
Co-Authored-By: jannikgro <[email protected]>

* delete some files

* bcolors back

* readme websocket

* debug websocket

* fixed websocket

* small websocket fix

* logging for websocket

* more logging?

* logging

* logging relevant content

* reset?

* reset webserver

* relevant gitignore stuff

* some resets in docker manager

* init files again?

* separate_markets again

* remove unnecessary code

* fixed docker tests, introduce Mocked logger

* logger into DockerManager

* Fixed typo bug

* log to file when executing `docker_manager`

Co-authored-by: Nikkel Mollenhauer <[email protected]>
Co-authored-by: jannikgro <[email protected]>
Co-authored-by: Johann Schulze Tast <[email protected]>
felix-20 added a commit that referenced this pull request Jul 30, 2022
* merge multiple experiments

commit 8652a67
Author: Judith <[email protected]>
Date:   Mon Apr 11 18:17:20 2022 +0200

    missed `handle_requests` on last commit

commit a354f89
Author: Judith <[email protected]>
Date:   Mon Apr 11 18:16:15 2022 +0200

    implement #356

commit fd54365
Author: Judith <[email protected]>
Date:   Mon Apr 11 17:56:35 2022 +0200

    docstrings

commit 9f43c9d
Author: Judith <[email protected]>
Date:   Mon Apr 11 15:07:47 2022 +0200

    fix docker test

commit b3e6e80
Merge: 6f64f9e 091eee6
Author: felix-20 <[email protected]>
Date:   Mon Apr 11 14:31:39 2022 +0200

    Merge branch 'development' into 379-multiple-experiments

commit 6f64f9e
Author: Judith <[email protected]>
Date:   Mon Apr 11 14:20:47 2022 +0200

    clean up docker port

commit 8c35e8a
Author: Judith <[email protected]>
Date:   Mon Apr 11 11:33:30 2022 +0200

    merge development

    commit 091eee6
    Author: Nikkel Mollenhauer <[email protected]>
    Date:   Mon Apr 11 10:43:00 2022 +0200

        Tests for `config_validation.py` (#404)

        * Don't get default data (shouldn't be necessary)

        * Unpack default data

        * Refactored utils functions to return dicts instead of strings

        * Adapted to new mock format

        * Adapted to new mock format

        * Some first tests

        * More tests

        * Fixed testcase-names

        * Moved file-endings to initial function call

        * Fixed tests

        * More asserts

        * More tests

        * More tests

        * More tests

        * `validate_sub_keys`-tests

    commit f8bc162
    Author: jannikgro <[email protected]>
    Date:   Fri Apr 8 13:20:42 2022 +0200

        [D] Stable baselines integration (#384)

        * refactored reinforcement learning agent to accept marketplace

        * adapted test_exampleprinter.py to marketplace initialization

        * add market option to accept continuos actions

        * fixed action space check

        * initial stable baselines integration

        * Agent init by env (#390)

        * introduced self.network in actorcritic_agent

        * added network_architecture in QLearningAgent

        * changed actorcritic_agent to network_architecture

        * set back training_scenario

        * am_configuration initialize rl-agent via marketplace

        * added final analyse to stable baselines training

        * added more stable baselines algorithms

        * added ppo algorithm

        * introduced stable_baselines_folder

        * renamed training to callback

        * satisfied linter

        * fixed loading problem

        * try to make tqdm run in stable_baselines

        * make tqdm running

        * reduced pmonitoring episodes in sb training

        * save model only if significantly better

        * fixed too long test time bug

        * moved back to 250 episodes testing

        * set timeout to 15 minutes

        * added first batch of fixes to @NikkelM feedback

        * added type annotations and asserts in stable_baselines_model

        * added sbtraining to training_scenario

        * applied comments in am_configuration

        * solved .dat problem and fixed crashing asserts

        * reintroduced _end_of_training

        * removed deprecated if

        * Moved '.dat' to function call instead of appending within function

        * Fixed assert

        * fixed model file ending bug

        * Add short explanation docstring

        Co-authored-by: Johann Schulze Tast <[email protected]>

        * fixed wrong docstring

        * Fixed tests

        Co-authored-by: NikkelM <[email protected]>
        Co-authored-by: Johann Schulze Tast <[email protected]>

commit 0183876
Author: Judith <[email protected]>
Date:   Mon Apr 11 11:31:15 2022 +0200

    name from `names` for container

commit ad70b7c
Author: Judith <[email protected]>
Date:   Sun Apr 10 16:32:47 2022 +0200

    fix #380

commit 8e1314c
Author: Judith <[email protected]>
Date:   Sun Apr 10 16:05:28 2022 +0200

    multiple experiments are supported on the webserver

commit 53a4000
Author: Judith <[email protected]>
Date:   Fri Apr 8 14:36:44 2022 +0200

    support for starting multiple container on docker side

commit 3c94f43
Author: Judith <[email protected]>
Date:   Fri Apr 8 11:46:07 2022 +0200

    ability to add `DockerInfo` for multiple container

* first attempt for websocket

* websocket on docker site working

* just send all of it if things have changed

* started on webserver site

* webserver sends push notification to user about stopped container

* commit

* websocket to ssl

* fix

* try

* started on database manager

* more db

* fix?

* debug statements

* more types

* does table exist fix

* colorful terminal output

* commit

* debug

* extra health checker

* debug for force stop

* some logging

* fix tests

* telegram notifications

* better logging

* better logging

* fix test?

* different logging level

* small fixes

* started on system monitoring

* system monitoring

* csv possibility for system data

* two buttons in webserver

* more fail prove

* fix configuration form

* silent_starter

* logging statements for silent starter

* debug

* debug

* more debugging

* fix error

* more debugging

* more local monitoring

* adjusted logging in silent starter

* gpu

* os.system

* subprocess

* Squashed commit of the following:

commit af12e13
Merge: bc9d3a3 ab704ca
Author: Nikkel Mollenhauer <[email protected]>
Date:   Tue Jun 7 18:20:56 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit bc9d3a3
Author: Nikkel Mollenhauer <[email protected]>
Date:   Tue Jun 7 18:20:50 2022 +0200

    Removed dead methods

commit ab704ca
Author: felix-20 <[email protected]>
Date:   Tue Jun 7 16:55:34 2022 +0200

    improved prefill (#496)

    * improved prefill with consideration of the current formdata

    * remove debug statements

    * remove print debug

    * Added lxml dependency

    * Removed debug comment

    Co-authored-by: Nikkel Mollenhauer <[email protected]>

commit 8839f90
Author: Judith <[email protected]>
Date:   Tue Jun 7 12:01:43 2022 +0200

    remove `config_is_final`

commit 53d79f9
Author: NikkelM <[email protected]>
Date:   Tue Jun 7 10:41:51 2022 +0200

    Review feedback by @SinNeax

commit 99ab681
Author: NikkelM <[email protected]>
Date:   Sat Jun 4 15:02:34 2022 +0200

    Fixed config validation

commit 6a09267
Merge: 179e0da e623413
Author: Judith <[email protected]>
Date:   Fri Jun 3 14:50:15 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit 179e0da
Author: Judith <[email protected]>
Date:   Fri Jun 3 14:49:49 2022 +0200

    fix javascript error

commit e623413
Author: NikkelM <[email protected]>
Date:   Fri Jun 3 14:38:20 2022 +0200

    Added `config_type` field to default modelfiles

commit bd15108
Merge: dac92d3 507cbc1
Author: NikkelM <[email protected]>
Date:   Fri Jun 3 09:45:11 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit dac92d3
Author: NikkelM <[email protected]>
Date:   Fri Jun 3 09:44:49 2022 +0200

    Fixed tests

commit 507cbc1
Author: Judith <[email protected]>
Date:   Fri Jun 3 07:52:30 2022 +0200

    dynamic table for sim_market possible

commit bce03bd
Author: Judith <[email protected]>
Date:   Thu Jun 2 20:15:17 2022 +0200

    fix webserver tests

commit 57dbb33
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 14:29:40 2022 +0200

    Webserver format

commit b3780cd
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 14:15:01 2022 +0200

    Fixed validation

commit c9657a9
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 13:50:20 2022 +0200

    Added back needed functionality of validating "complete" configs

commit dca24f3
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 13:34:21 2022 +0200

    Added some tests

commit fff9b47
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 13:12:41 2022 +0200

    Renamed `market` to `sim_market`

commit caf5c01
Merge: b8c43b7 f50fe65
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 11:30:18 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit b8c43b7
Author: Nikkel Mollenhauer <[email protected]>
Date:   Thu Jun 2 11:29:34 2022 +0200

    New config format

commit f50fe65
Author: Judith <[email protected]>
Date:   Thu Jun 2 11:21:10 2022 +0200

    implement new config validation for webserver

commit 135e8f3
Author: Judith <[email protected]>
Date:   Wed Jun 1 20:52:12 2022 +0200

    fix webserver tests

commit d326245
Author: Judith <[email protected]>
Date:   Wed Jun 1 17:26:50 2022 +0200

    fix prefill

commit 70492c8
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 17:08:34 2022 +0200

    Fixed remaining tests(?)

commit 59136b1
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 16:35:00 2022 +0200

    Fixed most tests

commit c912309
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 16:11:20 2022 +0200

    Removed references to `class`-field

commit 2ccee78
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 15:12:56 2022 +0200

    Removed `class` keywords from config files

commit c51dab7
Author: Nikkel Mollenhauer <[email protected]>
Date:   Wed Jun 1 15:09:46 2022 +0200

    renamed rl_config to q_learning_config

commit f9aa012
Author: NikkelM <[email protected]>
Date:   Tue May 31 19:49:14 2022 +0200

    Fixed Agent_monitoring with stable_baselines agents

commit 16df8c4
Merge: 7610c6e 26cb54b
Author: Judith <[email protected]>
Date:   Tue May 31 13:50:24 2022 +0200

    Merge branch '484-configuration-remove-combined-config' of https://github.com/hpi-epic/BP2021 into 484-configuration-remove-combined-config

commit 7610c6e
Author: Judith <[email protected]>
Date:   Tue May 31 13:50:14 2022 +0200

    script for creating rl model

commit 26cb54b
Author: NikkelM <[email protected]>
Date:   Tue May 31 11:21:46 2022 +0200

    Simplified and extended key validation

commit 0fa7eb5
Author: Judith <[email protected]>
Date:   Mon May 30 14:17:22 2022 +0200

    rl config works in view

commit da05474
Author: NikkelM <[email protected]>
Date:   Fri May 27 13:37:39 2022 +0200

    New feather website

commit dd40104
Author: NikkelM <[email protected]>
Date:   Fri May 27 12:56:28 2022 +0200

    Removed debug comment, fixed policyanalyzer

commit 1d025c9
Author: NikkelM <[email protected]>
Date:   Fri May 27 12:39:17 2022 +0200

    Reintroduced `main` for testing purposes

commit c90a8d6
Author: NikkelM <[email protected]>
Date:   Fri May 27 12:26:10 2022 +0200

    Fixed config validation for webserver

commit 4107aaf
Author: NikkelM <[email protected]>
Date:   Fri May 27 10:57:04 2022 +0200

    Added some debugging to docker

commit 2588f6d
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 22:22:39 2022 +0200

    preconditions for actor critic parameters and stable baselines parameter have been created

commit 4e174e6
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:42:23 2022 +0200

    restored changes

commit 14f181d
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:37:23 2022 +0200

    restored last changes

commit 3d3ac9d
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:34:11 2022 +0200

    more information

commit 31a212f
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:31:31 2022 +0200

    more information

commit 1b139e0
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:28:02 2022 +0200

    test

commit 1e80e8c
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:25:16 2022 +0200

    next try

commit 09a433e
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:22:45 2022 +0200

    another printf try

commit de9f5eb
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:19:07 2022 +0200

    more printf debug

commit 511d38a
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:16:11 2022 +0200

    printf debug

commit 1d1a3c4
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:12:06 2022 +0200

    restored last changes

commit 61c0d8e
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:09:22 2022 +0200

    tried without reading check

commit 068abaa
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:03:54 2022 +0200

    tried with dirty fix

commit f1cdf45
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 19:02:03 2022 +0200

    debug print

commit 75a674e
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 18:56:08 2022 +0200

    restored docker_manager

commit 39bc238
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 18:51:31 2022 +0200

    tried with assert False

commit 6986513
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 18:21:59 2022 +0200

    small change in docker_manager

commit 561fc00
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 18:13:18 2022 +0200

    tried to fix config_validation

commit 3fded29
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 17:25:10 2022 +0200

    reintroduced test_hyperparameter_config_market

commit b6a4945
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 15:52:19 2022 +0200

    reintroduced test_hyperparameter_config_rl

commit 7cef3e7
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 15:19:29 2022 +0200

    added rules and verifications

commit 473ddf7
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 12:43:12 2022 +0200

    introduced JSONConfigurable class and demanded class entry in config

commit 36024c8
Author: Jan Niklas Groeneveld <[email protected]>
Date:   Thu May 26 11:50:48 2022 +0200

    solved #484 without check

* automatically write template files

* try fix pre-commit

* try fix pre-commit

* try except

* more debug

* more debug

* debugging :)

* hotfix for qlearning file

* adopting to the given agent works again

* restore docker manager and app.py

* try restore again

* webserver tests are running again

merge from dev was very very strange

* new rl_config

* precommit?

* precommit!

* assert print

Co-Authored-By: Nikkel Mollenhauer <[email protected]>

* remove print

Co-Authored-By: Nikkel Mollenhauer <[email protected]>

* new regex for matching ce agents

Co-Authored-By: Nikkel Mollenhauer <[email protected]>

* fixed fixed price agent

Co-Authored-By: Nikkel Mollenhauer <[email protected]>

* comment out assert

Co-Authored-By: Nikkel Mollenhauer <[email protected]>
Co-Authored-By: jannikgro <[email protected]>

* delete some files

* bcolors back

* readme websocket

* debug websocket

* fixed websocket

* small websocket fix

* logging for websocket

* more logging?

* logging

* try make container_heath_checker only run when there are actually container running

* right keyword

* cls

* debug printing

* try popen

* debug print

* make it work again?

* should run monitoring for docker manager

* make webserver work again

* docstrings

* fix tests

* remove files add docstrings

* remove more files

Co-authored-by: Nikkel Mollenhauer <[email protected]>
Co-authored-by: jannikgro <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
discuss To be discussed research A feature offering value to Alex
Projects
None yet
Development

Successfully merging this pull request may close these issues.

[RL] Integrate DDPG from Stable-Baselines
3 participants