Release notes

Dev

New features:

  • The “simple” format can now include flags.

  • A max_records parameter, by default 10,000, has been added, to limit the size of uploaded time series.

Minor improvements:

  • A pandas FutureWarning has been removed.

  • Added some more log messages at the “info” level.

Version 3.0 (2022-12-05)

Briefly, this version is timezone-aware.

Enhydris 4.0 handles time zones better than 3.x, but it has some breaking changes in the API, namely that when uploading data the time zone must be specified. Accordingly, loggertodb 3.0 is compatible with Enhydris 4.0 and incompatible with previous Enhydris versions.

In this loggertodb version, the timezone configuration parameter must always be specified, and it is used to interpret all timestamps (whereas in previous versions it was used only for DST changes).

Version 2.2 (2021-01-13)

This makes loggertodb compatible with a new version of Enhydris that no longer has a “raw” time series type, and instead has an “initial” time series type. (This update has occurred in an internal unnumbered Enhydris release that precedes the release of Enhydris 3).

Changes in 2.2 microversions

2.2.1 (2022-11-06)

  • Bug fix: If an error occurs during data uploading, the error is correctly shown and logged (and the traceback also logged in DEBUG level), whereas in 2.2.0 it crashed with an irrelevant message.

Version 2.1 (2020-11-17)

The nullstr parameter has been renamed to null, keeping nullstr as a deprecated synonym. If the value of null is a number, it is interpreted as a number instead of as a string.

Version 2.0 (2020-10-14)

loggertodb version 2 is not compatible with Enhydris versions earlier than 3. Use loggertodb version 1 for Enhydris version 2.

Token authentication

loggertodb used to use a username and password to logon to Enhydris. Now it uses a token instead. Accordingly, the username and password configuration file parameters have been abolished and auth_token has been introduced. See the documentation on authentication for more detailed instructions.

Time series groups

Enhydris 3 contains the notion of a time series group—a group of related time series for the same variable, e.g. the initial, checked and aggregated versions of a time series. Accordingly, loggertodb has been changed so that the fields parameter and the wdat5-specific meteorological parameters specify time series group ids rather than time series ids. loggertodb will always upload data in the “initial” time series of the time series group; if such a time series does not exist, it is automatically created.

How to upgrade from version 1

If you have a loggertodb v1 configuration file, loggertodb v2 can convert it. Enter this:

loggertodb --upgrade loggertodb.conf

where loggertodb.conf is the configuration file you want to upgrade. loggertodb will make requests to Enhydris in order to determine the auth_token from username and password, and the time series group ids from the time series ids. The configuration file will be upgraded accordingly. The original file will be backed up by adding the .bak extension (e.g. loggertodb.conf.bak). (If the backup file already exists and is different from the original, loggertodb will terminate with an error.)

Changes in 2.0 microversions

2.0.1 (2020-10-21)

  • Fixed inability of the 2.0.0 Windows executable to run on Windows 7.

2.0.2 (2020-10-23)

  • Fixed malfunctioning Windows executable (it had been built with wrong dependencies).

History up to Version 1

1.0.0 (2019-10-29)

  • Improved handling of switch from DST to winter time.

0.2.2 (2019-08-20)

  • Improved error message in multi-file simple format when nfields_to_ignore was 1 or more and a line did not have enough fields.

0.2.1 (2019-07-17)

  • Fixed a crash when a file was empty in multi-file simple format.

  • Improved error messages in multi-file simple format when the timestamps were badly ordered in a file or overlapping between files.

0.2.0 (2019-07-16)

  • Added multi-file option to simple format.

  • Added configuration parameters “encoding” and “ignore_lines”.

0.1.3 (2019-06-07)

  • Upgraded htimeseries to 1.0.

  • Made dependencies more robust.

0.1.2 (2019-05-27)

  • Made parsing dates more robust in simple format.

  • Fixed extreme slowness when thousands of records had to be inserted.

  • Fixed unhelpful error message when file was out of order.

0.1.1 (2019-04-18)

  • Fixed a bug that prevented using a log file.

0.1.0 (2019-04-18)

  • Initial release