Changelog#

Unreleased#

[0.4.13] - 2026-02-11#

Added#

  • Add keep_age, keep_sex, and keep_starttime parameters to Edf.anonymize() for optionally keeping these fields when anonymizing an EDF/BDF record (#100 by Clemens Brunner).

[0.4.12] - 2026-01-20#

Changed#

  • Avoid loading data from disk when creating a new Edf object based on signals from an existing file (#98).

[0.4.11] - 2025-11-19#

Fixed#

  • Use lazy loading in Edf.slice_between_seconds to speed up execution (#93).

[0.4.10] - 2025-10-07#

Added#

  • Support reading and writing BDF and BDF+C files (#63).

Fixed#

  • Changing the microsecond part of an EDF+C with multiple annotation signals now correctly updates all annotation onsets (#63).

[0.4.9] - 2025-05-19#

Fixed#

  • When reading a file where the number of data records is set to -1 in the header, automatically fix the corresponding field and issue a warning (#74).

[0.4.8] - 2025-03-26#

Added#

  • Add a parameter header_encoding="ascii" to read_edf() to support reading header fields that use a character encoding different to ASCII (#68 by JohnAtl).

Fixed#

  • Use ASCII instead of UTF-8 as the default character encoding for reading header fields (#68 by JohnAtl).

[0.4.7] - 2025-03-21#

Fixed#

  • Allow having annotations with and without durations in the same recording (#67 by JohnAtl).

[0.4.6] - 2025-02-27#

Fixed#

  • Improve efficiency of creating annotations signals in EDF+ files (#65).

[0.4.5] - 2024-10-28#

Added#

  • Load a segment of EDF raw data from a signal without reading the entire file using EdfSignal.get_data_slice method (#58).

  • Load annotations from a specific time window of the EDF file without reading the entire file using Edf.get_annotations method (#58).

[0.4.4] - 2024-09-25#

Added#

  • Allow adding new EDF+ annotations and overwriting all existing ones with Edf.add_annotations and Edf._set_annotations (#54).

[0.4.3] - 2024-06-23#

Fixed#

  • Retrieving data for signals where physical minimum and maximum are the same, the warning message now correctly reflects that (#51).

[0.4.2] - 2024-05-13#

Added#

  • Lazy load raw signals from local EDF files. Lazy loading is supported and enabled by default when passing a str or pathlib.Path object to the read_edf function. The behavior can be controlled using the lazy_load_data parameters of the read_edf function (#48).

  • Public getter for digital signal values called digital (#48).

[0.4.1] - 2024-04-30#

Added#

  • Make EDF+ header fields patient and recording more tolerant regarding non-compliant inputs: omitted subfields are returned as X instead of throwing an exception (#18).

  • Allow reading from tempfile.SpooledTemporaryFile[bytes] (#36).

  • Enable the reading of EDF files where the filesize does not match the header information. Incomplete data records will be truncated (#41).

Fixed#

  • Allow reading more non-standard starttime and startdate fields (#30).

[0.4.0] - 2023-12-22#

Changed#

  • Exclude EDF+ annotation signals from Edf.signals, Edf.num_signals, and Edf.drop_signals() (#25).

  • Provide more concise __repr__ for Edf and EdfSignal (#26).

  • Edf.append_signals() now inserts new signals after the last ordinary (i.e. non-annotation) signal (#29).

Added#

  • Expand ~ (tilde) to the user’s home directory in edf_file argument of read_edf() and target argument of Edf.write() (#23).

Fixed#

  • Avoid floating point errors sometimes preventing the creation of an Edf with signals that are actually compatible in duration (#15).

  • Allow reading EDF+ startdate and birthdate with non-uppercase month (#19).

  • Disallow setting signal label to "EDF Annotations" (#28).

[0.3.1] - 2023-12-06#

Added#

  • Add optional parameter sampling_frequency to EdfSignal.update_data() for changing the sampling frequency of the signal when updating its data (#13).

  • Add update_data_record_duration method to class Edf for updating the data_record_duration field of the EDF header (#14).

[0.3.0] - 2023-11-30#

Changed#

  • Edf.signals can not be set anymore. Use Edf.append_signals(), Edf.drop_signals(), and EdfSignal.update_data() instead (#10).

Added#

  • Allow creating a new Edf containing only annotations (#7).

  • Add EdfSignal.update_data for overwriting the physical values with an array of equal length (#10).

  • Allow adding new signals to an Edf with Edf.append_signals (#10).

Fixed#

  • Disallow creating a new Edf where local patient/recording identification subfields are empty strings (#6).

  • Allow retrieving the starttime from a file where the reserved field indicates it is an EDF+C but no annotations signal is present (#8).

  • Disallow removing the EDF+ timekeeping signal with Edf.drop_signals (#10).

[0.2.1] - 2023-11-17#

Fixed#

  • Disallow creating a new Edf with a recording startdate not between 1985-01-01 and 2084-12-31 (#5).

[0.2.0] - 2023-11-16#

Added#

  • Support non-standard header fields (not encoded as UTF-8) by replacing incompatible characters with “�” (#4).

Fixed#

  • When EdfSignal.physical_min or EdfSignal.physical_max do not fit into their header fields, they are now always rounded down or up, respectively, to ensure all physical values lie within the physical range (#2).

  • The calculation of num_data_records from signal duration and data_record_duration is now more robust to floating point errors (#3).

[0.1.1] - 2023-11-09#

Fixed#

  • Use correct path to sphinx config file in .readthedocs.yml.

[0.1.0] - 2023-11-09#

Initial release 🎉

Added#

  • Support for reading and writing EDF and EDF+C files.

  • I/O from/to files and file-like objects.

  • Edf.slice_between_seconds() and Edf.slice_between_annotations() for slicing a recording in time.

  • Edf.drop_annotations() for dropping EDF+ annotations based on their text.

  • Edf.drop_signals() for dropping individual EDF signals by specifying their index or label.

  • Edf.anonymize() for anonymizing all identifying header fields.