-
Notifications
You must be signed in to change notification settings - Fork 44
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
Warm start a stratey based on config provided seed conditions #487
base: main
Are you sure you want to change the base?
Conversation
Summary: Update db master table spec such that experiment ID and participant is not unique. Functions that used experiment_id as a key have been swapped to use the master table's master key (unique_id). Functions appeared to expect that it's possible to get multiple entries back from a single experiment ID existed but didn't make sense since it was unique, this means nothing really needs to be changed. Replay functions used to leverage the experiment_id, here it is assumed to be unique, which it was. Now replay functions use the master table's unique_id to pick which experiment to replay, which conveniently also means it's a lot easier to just try integers starting from 0 (instead of finding a uuid). Metadata reading has been changed to correctly get all the information from the config to match the master table spec. These are technically breaking changes that may affect old scripts but it's not clear what db utility functions may be used in weird scripts trying to use experiment_ids to identify experiments. Still old dbs should all work and be compatible. Differential Revision: D66526187
This pull request was exported from Phabricator. Differential Revision: D66731597 |
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same Differential Revision: D66731597
21dd86e
to
c0016b9
Compare
This pull request was exported from Phabricator. Differential Revision: D66731597 |
1 similar comment
This pull request was exported from Phabricator. Differential Revision: D66731597 |
c0016b9
to
c9ee936
Compare
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same Differential Revision: D66731597
This pull request was exported from Phabricator. Differential Revision: D66731597 |
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same Differential Revision: D66731597
c9ee936
to
90efa22
Compare
This pull request was exported from Phabricator. Differential Revision: D66731597 |
90efa22
to
0bf8ed4
Compare
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same Differential Revision: D66731597
0bf8ed4
to
a361adb
Compare
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same allow-large-files Differential Revision: D66731597
This pull request was exported from Phabricator. Differential Revision: D66731597 |
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same allow-large-files Differential Revision: D66731597
a361adb
to
0969773
Compare
This pull request was exported from Phabricator. Differential Revision: D66731597 |
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same allow-large-files Differential Revision: D66731597
0969773
to
a78ee13
Compare
This pull request was exported from Phabricator. Differential Revision: D66731597 |
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same allow-large-files Differential Revision: D66731597
This pull request was exported from Phabricator. Differential Revision: D66731597 |
a78ee13
to
9d35c6f
Compare
This pull request was exported from Phabricator. Differential Revision: D66731597 |
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same allow-large-files Differential Revision: D66731597
9d35c6f
to
040ebc1
Compare
…okresearch#487) Summary: Pull Request resolved: facebookresearch#487 There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments. Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria: - experiment_name - experiment_description - experiment_id - participant_id - anything stored in extra_metadata Criteria will follow AND logic between fields, but inclusive OR logic within the same field. Data will be further filtered out if it does not meet these criteria: - Each parameter matches on name and type - Each outcome matches on name and type - Stimuli per trial are the same allow-large-files Differential Revision: D66731597
This pull request was exported from Phabricator. Differential Revision: D66731597 |
040ebc1
to
b53af2f
Compare
Summary:
There is a desire to warm start a strategy with data filtered on master table values recorded from previous experiments.
Data can be filtered on any data other than master.unique_id. This means that the below are all valid filter criteria:
Criteria will follow AND logic between fields, but inclusive OR logic within the same field.
Data will be further filtered out if it does not meet these criteria:
Differential Revision: D66731597