attempt txt input update

[sanghoonio]

This is to allow create_igd_f() to work with a .txt input file instead of a directory for fileList of beds.

[nleroy917]

Oh so this will basically say:

If filelist is an actual file on disk, read it in... otherwise assume its a directory and loop through

[nleroy917]

I feel like it should be better to just have it read from stdin, right? That way you can be like:

cat filelist.txt | gtars igd

or

ls path/to/files/*.bed.gz | gtars igd

[sanghoonio]

For usability yeah, just added the txt file input for now because it seemed easier to add. I'll try stdin inputs

[donaldcampbelljr]

I feel like it should be better to just have it read from stdin, right? That way you can be like:

cat filelist.txt | gtars igd

or

ls path/to/files/*.bed.gz | gtars igd

The C version builds from a directory or a .txt file which is why I suggested it (to continue towards total feature parity). Is reading from stdin objectively better?

[sanghoonio]

From an R bindings perspective having to write a txt file from R to read in rust might negligibly slow things down over passing in bed filepaths directly. With stdin though, we can input multiple paths like below right?

path/to/files/file1.bed.gz path/to/files/file2.bed.gz path/to/files/file3.bed.gz | gtars igd

[donaldcampbelljr]

From an R bindings perspective having to write a txt file from R to read in rust might negligibly slow things down over passing in bed filepaths directly. With stdin though, we can input multiple paths like below right?

path/to/files/file1.bed.gz path/to/files/file2.bed.gz path/to/files/file3.bed.gz | gtars igd

Ah ok. Got it. Thanks.

[nleroy917]

The C version builds from a directory or a .txt file which is why I suggested it (to continue towards total feature parity). Is reading from stdin objectively better?

I'm not too sure which is better. I think its more in line with "UNIX philosophy" to be able to pipe things around, but I suppose I'm not too opinionated one way or another.

[donaldcampbelljr]

I fixed this and the functionality now works for reading a .txt file.

[donaldcampbelljr]

And I added stdin as an option so you can do this:

echo "/home/drc/GITHUB/gtars/gtars/tests/data/igd_file_list/igd_bed_file_1.bed" | cargo run igd create --output /home/drc/IGD_TEST/output/ --filelist -

And it works.

Please confirm this meets your needs and then merge to dev when you are ready.

[donaldcampbelljr]

Just make sure this works like you expect it to.

[sanghoonio]

And I added stdin as an option so you can do this:

echo "/home/drc/GITHUB/gtars/gtars/tests/data/igd_file_list/igd_bed_file_1.bed" | cargo run igd create --output /home/drc/IGD_TEST/output/ --filelist -

And it works.

Please confirm this meets your needs and then merge to dev when you are ready.

Looks like it creates a .igd with .txt input but it the .igd output doesn't seem to work with search. A .igd created with a bed file directory doesn't work with search either unless I switch back to dev and recompile. Can you let me know if you can reproduce this?

sam@Sams-MacBook-Pro release % ./gtars igd create --output /Users/sam/Documents/Work/episcope/.test/igd/ --filelist /Users/sam/Documents/Work/episcope/.test/bed_paths_10.txt

Number of Bed Files found:
10
Temporary Tiles:
 nCtgs (igd.nctg): 63, nRegions (igd.total): 4840383, nTiles (nt): 191146
IGD saved to: /Users/sam/Documents/Work/episcope/.test/igd/igd_database.igd
Total Intervals: 4644521, l_avg: 690.1135
nctg:63  nbp:16384
sam@Sams-MacBook-Pro release % ./gtars igd search --database /Users/sam/Documents/Work/episcope/.test/igd/igd_database.igd --query /Users/sam/Documents/Work/episcope/.test/bed2.bed

thread 'main' panicked at src/igd/search.rs:434:41:
index out of bounds: the len is 0 but the index is 0
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

[donaldcampbelljr]

I can't reproduce the above error. However, when using the dev branch of IGD, I notice a bug where I'm creating an .igd from a directory and my search is returning 0 hits.

However, if I create an igd from a single source file and query that source file against the igd, it appears to be working fine.

I will have to investigate this early next week. I suspect the issues may be related.

[sanghoonio]

Hmm I'm still getting the same error after pulling and recompiling, are there differences between mac and linux when compiling rust that can cause something like this?