There is a problem with the splitting function

At the moment test only works when build in the same directory as the code is, because test uses the path ./example

Use robin_hood::unordered_set instead of std::unordered_set

robin_hood::unordered_node_map instead of robin_hood::unordered_map?

Inserting new experiments needs to be implemented

Probably because it is a Git submodule.

Add cli tests as proposed by https://github.com/joergi-w/app-template/wiki

See here for why: https://github.com/google/googletest/blob/master/googletest/docs/faq.md#why-should-test-suite-names-and-test-names-not-contain-underscore

Add threshold option to search, to make it more comparable to other tools like Mantis.

It might make sense to use a different normalization method after preprocessing, this should be possible for all methods based on a genome mask, where new sequences are given. Not possible remains to calculate anything on the sequences of the experiment, because these are not given.

Instead of having one fpr for all files on a level, estimate the fpr for every file, store it and then use it in the estimations. This is important, if the amount of file content differs a lot, because then the fpr for all is not correct.

Using a threshold not 0.5 in the search returns a cereal error.

In order to gurantee a fpr for every file, I could change from using the mean the maximum... This would increase the size, but compression should be even smaller.

Currently, the header file information is not used at all, when using ibf with minimiser files. So, an user would have to type in cutoffs, k, w, shape again. This does not make sense...
On the other hand, the expression level information should be not used, there is no reason, why an user wants to stay with those. They should only be used for statistics.

Maybe the information about k, w, shape can be stored in a better way, so that not an extra file is necessary? Like in the name? experiment_k_w_shape_cutoff.minimiser could work. Then the header would only be used for the statistic function.
Or stome them in a data structure.

If the genome file is given, the mean normalization works still on all minimizers.

Sometimes the ape test test_needle fails for the multithread option.

[ RUN      ] ibfmin.no_given_expression_levels_multiple_threads
/home/mitradarvish/Dokumente/develop/needle/test/api/test_needle.cpp:337: Failure
Expected equality of these values:
  expected_result2
    Which is: [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1]
  res2
    Which is: [1,0,1,1,1,1,1,1,1,1,1,0,1,1,1,1]
[  FAILED  ] ibfmin.no_given_expression_levels_multiple_threads (1 ms)

• Add HIBF #100
• Add Merge option #38
• Add update option? #38
• Make Preprocessing better with multiple threads
• Make count competitive with kallisto & Co.?
• Add argument verbose #32

For estimating differential expressed genes, add a normalization method inspired by the DESeq algo.
(see: https://hbctraining.github.io/DGE_workshop/lessons/02_DGE_count_normalization.html)

At the moment the mantis cutoffs are only used when the minimiser function is called beforehand.

Add a possibility to add a new experiment and a possibility to add an IBF with a new expression level. For adding a new expression level, the IBFs "cornering" the new expression value need to be recalculated.

I'll do a PR that:

• Updates SeqAn to 3.2.0 (#126)
• Updates robin-hood (#126)
• Integrates sharg-parser
• Updates CI (#126)

[ ] hibf_lib
[ ] chopper

I put together a basic package of Needle for Debian and I discovered that there is another programs with the same name

/use/bin/needle is already provided by EMBOSS
https://packages.debian.org/bullseye/emboss a.k.a https://bio.tools/needle-ebi or https://bio.tools/needle

This causes a problem with Debian policy

Two different packages must not install programs with different functionality but with the same filenames. […] If this case happens, one of the programs must be renamed.

https://www.debian.org/doc/debian-policy/ch-files.html#binaries

And I would rather that the Debian package of Needle has the same program name as other packaging systems. Otherwise that causes problems for user's scripts and workflows..

Perhaps this Needle could be renamed? Sorry to ask it, but it would be best to not have the confusion.

Add an optional preprocessing step that counts all minimizers and their occurrences for an experiment and saves it in a binary file, the ibf should then be build on these binarys.

Add command line argument verbose

Rename genome mask to include.
Add option to exclude certain minimisers.

https://buildd.debian.org/status/fetch.php?pkg=seqan-needle&arch=amd64&ver=1.0.2%2Bds-2&stamp=1704746033&raw=0

• Add link to apptemplate in Readme
• Explain how doc can be built
• Check documentation for errors or missing paragraphs
• Add more information, so that built documentation is easier to use. Maybe add a small tutorial?