Quickstart¶

Initialize DB¶

import pymisha as pm

pm.gdb_init("/path/to/misha_db")

Extract values¶

intervals = pm.gintervals_from_strings(["chr1:0-1000", "chr1:2000-2600"])
out = pm.gextract("track1", intervals, iterator=100)

Screen and summarize¶

filtered = pm.gscreen("track1 > 0.5", intervals)
stats = pm.gsummary("track1", intervals)

Thread safety¶

PyMisha is not thread-safe. All state (active database, virtual tracks, config) is process-global, so you should not call PyMisha from multiple threads or open more than one database per process. See the README for full details.

Example DB¶

pm.gdb_init_examples()
print(pm.gtrack_ls())
print(pm.gextract("dense_track", pm.gintervals("chr1", 0, 1000)))

Seeing what PyMisha recovered from¶

PyMisha falls back rather than failing in a number of places - an optional dependency that is missing, a probe that decides a track is not of some type, a best-effort cleanup. Each fallback reports what it caught on the pymisha logger, which is silent until you configure logging:

import logging
logging.basicConfig()
logging.getLogger("pymisha").setLevel(logging.DEBUG)

Every module logs to its own child (pymisha.tracks, pymisha.intervals, ...), so a single one can be turned up on its own. Failures that R misha itself warns about are raised as pymisha.PymishaWarning instead, so they are visible with no configuration; silence them with warnings.filterwarnings("ignore", category=pm.PymishaWarning).