Closed GrowlingM1ke closed 7 months ago
Sounds great!!! Thank you for this update..- ym
From: Michal A Lange @.> Sent: Thursday, February 29, 2024 1:47 PM To: cci-MaLab/Cell_Clustering_Tool @.> Cc: Subscribed @.***> Subject: [External] [cci-MaLab/Cell_Clustering_Tool] Make Calculation of Missed Cell Signal more efficient. (Issue #4)
This message was sent from a non-IU address. Please exercise caution when clicking links or opening attachments from external sources.
Currently on a good machine. For a missed cell the calculation of the trace takes around 5-7 seconds. I've made it so that if the calculation was already done it won't be repeated. However, there is a way to make this approach significantly faster but it would require some refactoring:
— Reply to this email directly, view it on GitHubhttps://github.com/cci-MaLab/Cell_Clustering_Tool/issues/4, or unsubscribehttps://github.com/notifications/unsubscribe-auth/A3VZG6LYE5QQCKZ76PFITJ3YV53SVAVCNFSM6AAAAABEAO4YFKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGE3DCOBYGY4TKNY. You are receiving this because you are subscribed to this thread.Message ID: @.**@.>>
Fixed as of this commit: https://github.com/cci-MaLab/Cell_Clustering_Tool/commit/c8ff1b3a5490c1963ccd79eac4eff0f7b16488bb
Minian actually does save a rechunked version of the processed file called Y_hw_chk, which I replaced for Y_fm_chk. By indexing the sections associated with the mask values the Raw Signal gets extracted in less than a second as opposed to 30s to a minute as it was before.
Currently on a good machine. For a missed cell the calculation of the trace takes around 30 seconds. I've made it so that if the calculation was already done it won't be repeated. However, there is a way to make this approach significantly faster but it would require some refactoring: