htslib icon indicating copy to clipboard operation
htslib copied to clipboard

hts wide maxdepth limit check

Open vasudeva8 opened this issue 4 months ago • 0 comments

This is a draft, to fix https://github.com/samtools/samtools/issues/2238

htsFile updated to hold extra data, when max depth check is enabled. The max depth check is set using input-fmt-option, seq_max_depth. Extra data to hold are, buffer to track depth on a window of positions, tid, start/end of window and max depth above which reads can be discarded. This check is made only for data sorted on position and for others this shall not be used.

A buffer is allocated, which holds depth against positions in read. New reads update the depth, for all positions it spans, and if depth is below the limit for any position, it is passed to user and discarded otherwise. Read continues and another read covering a position with lower depth, if found is passed to user and read finishes with end of file if nothing else is found. As a new higher start position is found, depth up to that position is discarded and window is updated to hold data from that position onwards.

Pair reads are not considered in this.

vasudeva8 avatar Oct 30 '25 17:10 vasudeva8