SMLM Spacetime Resolution merge requestshttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests2022-11-18T00:04:43Zhttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/25Add tools for packaging as a toolbox2022-11-18T00:04:43ZThomas Rogers ShawAdd tools for packaging as a toolboxThree pieces:
smlm-analysis.prj is matlab's xml-based project specification file, with info such as what to add to path, author and version information
makepackage.m is an mfile that compiles the mex and then runs the packaging routine t...Three pieces:
smlm-analysis.prj is matlab's xml-based project specification file, with info such as what to add to path, author and version information
makepackage.m is an mfile that compiles the mex and then runs the packaging routine to produce an mltbx file
make_package.sh is a shell script that gets the version number according to git and then calls out to the m-filehttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/24Resolve "Very slow time_edge_correction for long datasets"2022-10-25T19:15:59ZThomas Rogers ShawResolve "Very slow time_edge_correction for long datasets"Closes #16
Three kinds of solution have been proposed:
1. If your data doesn't have gaps between chunks, we can use a much nicer algorithm to do the sum.
2. assume some simple (non-constant) functional form for the rate N(t) (e.g. exp...Closes #16
Three kinds of solution have been proposed:
1. If your data doesn't have gaps between chunks, we can use a much nicer algorithm to do the sum.
2. assume some simple (non-constant) functional form for the rate N(t) (e.g. exponential, or piecewise linear), and fit that to data, then do most of the work analytically.
3. initially use a simple approximation to figure out how to split the tau bins, and then only compute the complicated time edge correction for the wide tau bins that are used at the end.
(2) was rejected as irrelevant for this kind of data. It could be something nice to do later, but will not help in this case. For (1), I realized that this kind of algorithm is probably better anyway, and when there are N chunks with gaps, we can just do the N^2 convolutions and merge them. (3) is also quite doable, although not guaranteed to give the same answer, so maybe not the right thing to do by default.
So my plan of action is:
1. Write the code that would do a fast convolution, and test that it gives correct results for a single chunk.
2. Add the option to use the `'unif'` option for the acor that is used to select tau bins, but the `'actual'` option for the final results.
3. Explore how much work it would be to use the algorithm from 1 on N^2 pairs of chunks and then combine the results at the appropriate taus.https://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/23Resolve "LSF fits have artificial lower bound at 5 nm"2022-09-13T06:26:13ZThomas Rogers ShawResolve "LSF fits have artificial lower bound at 5 nm"Closes #17
Proposed fix: set the lower bound for the fit at the upper edge of the first radial bin of the autocorrelation.Closes #17
Proposed fix: set the lower bound for the fit at the upper edge of the first radial bin of the autocorrelation.https://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/22Draft: Resolve "Out of memory `spacewin_isinside` for polyshape type"2022-09-10T00:02:27ZThomas Rogers ShawDraft: Resolve "Out of memory `spacewin_isinside` for polyshape type"Closes #15
Issue was originally reported by @jcflan for spatial windows with many vertices.
Rough plan of action:
- get a rough sense of how many vertices and/or how many points in the dataset there need to be to cause a problem
- bre...Closes #15
Issue was originally reported by @jcflan for spatial windows with many vertices.
Rough plan of action:
- get a rough sense of how many vertices and/or how many points in the dataset there need to be to cause a problem
- break x and y into smaller chunks if they exceed this limit
- make it more convenient to convert between spacewin types
Look into additionally:
- possibly autoconvert to `'image'` spacewin type, unless otherwise specified?https://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/21Fix time_edge_correction_unif ...2022-09-09T23:44:23ZThomas Rogers ShawFix time_edge_correction_unif ...and add an option to use that time edge correction in spacetime_resolutionand add an option to use that time edge correction in spacetime_resolutionhttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/20Fig52022-08-29T21:07:46ZThomas Rogers ShawFig5Add figure 5 from the protocol chapterAdd figure 5 from the protocol chapterhttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/19Xcor protocol examples2022-08-29T20:58:33ZThomas Rogers ShawXcor protocol examplesAdd figure 5 as exampleAdd figure 5 as examplehttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/18Xcor protocol examples2022-08-29T20:30:52ZThomas Rogers ShawXcor protocol examplesBesides adding the figure scripts (slightly modified for clarity), the big change here is a revamped `README.md`, which now has brief descriptions of all of the important code, mentions the new examples, has a general overview paragraph,...Besides adding the figure scripts (slightly modified for clarity), the big change here is a revamped `README.md`, which now has brief descriptions of all of the important code, mentions the new examples, has a general overview paragraph, etc.https://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/17added function that corrects for long range spatial gradients for use in auto...2022-08-26T04:31:12ZSarah Veatchadded function that corrects for long range spatial gradients for use in auto-...added function that corrects for long range spatial gradients for use in auto- and cross-correlationadded function that corrects for long range spatial gradients for use in auto- and cross-correlationhttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/16Clean up some odds and ends2022-06-28T19:36:01ZThomas Rogers ShawClean up some odds and endsFix some help text.Fix some help text.https://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/15Added ability to do 2D LSFs (and autocorrelations in x,y,t)2022-06-28T19:30:31ZThomas Rogers ShawAdded ability to do 2D LSFs (and autocorrelations in x,y,t)Add code to do 2d (x,y) versions of the lsf calculation, as well as the autocorrelation that underlies it.Add code to do 2d (x,y) versions of the lsf calculation, as well as the autocorrelation that underlies it.https://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/14Move licence info below help text2022-06-08T17:19:18ZThomas Rogers ShawMove licence info below help textThis just makes it so matlab's `help` command shows the right
comment to the userThis just makes it so matlab's `help` command shows the right
comment to the userhttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/13Add spatial versions of acor and xcor2022-04-07T19:37:56ZThomas Rogers ShawAdd spatial versions of acor and xcorCloses #3
Closes #4
These basically are just the spacetime versions with the temporal factors taken out. I have tested these on some orai1 data.Closes #3
Closes #4
These basically are just the spacetime versions with the temporal factors taken out. I have tested these on some orai1 data.Jenny Flanagan-NatoliJenny Flanagan-Natolihttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/12Add ability to filter points based on timewin2022-03-31T19:14:36ZThomas Rogers ShawAdd ability to filter points based on timewinCloses #11
New function timewin_isinside returns inds of points that are inside
the timewin. This function is applied in the spacetime correlation
functions and spacetime_resolutions to filter points to the specified
temporal window.
...Closes #11
New function timewin_isinside returns inds of points that are inside
the timewin. This function is applied in the spacetime correlation
functions and spacetime_resolutions to filter points to the specified
temporal window.
The algorithm for timewin_isinside is probably not the absolute most efficient possible, but is quite fast (a ~10^5 point dataset took about .01s to run).Sarah VeatchSarah Veatchhttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/11This file is fake2022-03-31T20:17:21ZJenny Flanagan-NatoliThis file is fakehttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/10Documentation improvements in spacetime_{x,a}cor and spacetime_resolution2022-03-29T15:00:15ZThomas Rogers ShawDocumentation improvements in spacetime_{x,a}cor and spacetime_resolutionThis gets us to a reasonable amount of help text, improves some error messages, and gives dependency information in the readme.This gets us to a reasonable amount of help text, improves some error messages, and gives dependency information in the readme.https://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/9spacetime_xcor and closepairs interfaces2022-03-17T22:45:52ZThomas Rogers Shawspacetime_xcor and closepairs interfaces1. A new function `spacetime_xcor` that gives spatiotemporal cross-correlations
2. Improved interfaces to closepairs/crosspairs that decide for the user whether to sort on x or t.
There were also some bug fixes, especially in crosspairs...1. A new function `spacetime_xcor` that gives spatiotemporal cross-correlations
2. Improved interfaces to closepairs/crosspairs that decide for the user whether to sort on x or t.
There were also some bug fixes, especially in crosspairs, which had hard-to-debug indexing issues when `(t2 - t1 == taumax)`
Closes #7
Closes #5https://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/8Add LICENSE2022-03-17T16:04:42ZThomas Rogers ShawAdd LICENSEThis was left off earlierThis was left off earlierPolish for publicationhttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/7Resolve "Add license info"2022-03-13T02:31:44ZThomas Rogers ShawResolve "Add license info"Closes #13
Apply GPL v3 license.Closes #13
Apply GPL v3 license.Polish for publicationThomas Rogers ShawThomas Rogers Shawhttps://gitlab.umich.edu/veatchlab/smlm-spacetime-resolution/-/merge_requests/6Resolve "Option to return spacewin as output arg from spacewin_gui"2022-03-10T17:38:06ZThomas Rogers ShawResolve "Option to return spacewin as output arg from spacewin_gui"Closes #10
This is more or less as planned in the issue, although some details needed to be worked out.Closes #10
This is more or less as planned in the issue, although some details needed to be worked out.Polish for publicationThomas Rogers ShawThomas Rogers Shaw