tools to update target order through ptolemy active-learning score recalculation - Githubissues

Original Redmine Comment Author Name: Anchi Cheng (@anchi2c) Original Date: 2023-03-21T02:51:14Z

Tool 1 - Leginon update target order tool in MosaicLearnTargetFinder class. Typical node name-Square_Targeting Usage:

After auto square finder tool is run and target submitted, this tool is used to update the current target order. Clicking this the first time after target submitted update the order according to the pre-trained model. After that, it will update the order according to the active learning results and thus manually update the order instead of the auto process in looplearn.py described below.

Tool 2 - Appion looplearn.py

Usage: This appion loop script query database for ctf results and ALS ice thickness values of each image and send these results to Ptolemy to learn. It also keeps track of how many squares has been sent to Ptolemy and will auomatically update square target order.

The interface is accessed through myamiweb processing link similar to other appion loop.
This requires some ctf finding run has started.
Use aligned preset and wait for more images.
If desirable, adjust intevals of the automated target order to update.

leginon-org / leginon-redmine-archive

tools to update target order through ptolemy active-learning score recalculation #14509