WIP: Add support for reading and writing .wfdb archive files

[tompollard]

As discussed in wfdb/wfdb-spec#21 and #540 we would like to allow WFDB Records to be saved to a single file .wfdb.

This pull request introduces support for reading and writing .wfdb archive files, a ZIP-based format that stores .hea, .dat, and related files in a single container.

Implementation

• Adds WFDBArchive class to handle reading/writing of .wfdb archives
• Archives are implemented as uncompressed ZIP files for maximum simplicity
• Two ways to create archives:
  1. Inline creation via wrsamp(..., wfdb_archive=...) which writes files directly to the archive
  2. Two-step creation via wrsamp(...) followed by WFDBArchive.create_archive(...) which writes files to disk first
• Adds support for reading records directly from .wfdb archives via rdrecord()
• Adds tests covering:
  • Round-trip read/write of archives
  • Reading subset of channels
  • Reading partial samples
  • Error handling for missing files
  • Both inline and two-step archive creation methods

Example usage

import wfdb from wfdb.io.archive import WFDBArchive import numpy as np # 1. Read an existing WFDB record record = wfdb.rdrecord('sample-data/test01_00s') # 2. Convert physical signals to digital signals # Side note, but this shouldn't be necessary. # ...or at least the write error should be clearer. record.d_signal = record.adc() # 3. Write the record to the archive record.wrsamp(wfdb_archive="new_archive.wfdb") # 4. Read back from the archive archive_record = wfdb.rdrecord('new_archive.wfdb') # 5. Verify the data matches assert record.fs == archive_record.fs assert record.n_sig == archive_record.n_sig assert record.sig_name == archive_record.sig_name assert record.units == archive_record.units assert record.p_signal.shape == archive_record.p_signal.shape