v0.3.9

What's Changed

• [Fix] network id handling by @jthurner in #65

Full Changelog: v0.3.8...v0.3.9

v0.3.8

What's Changed

• enforce specifying variant argument as int instead of list by @jthurner in #61
• bump version to 0.3.8 by @jthurner in #62

Full Changelog: v0.3.7...v0.3.8

v0.3.7

What's Changed

• allow $or and $and in timeseries metadata query by @jkupka in #55
• Fix: pyproject requirements by @jthurner in #59
• bump to v0.3.7 by @jthurner in #60

New Contributors

• @jkupka made their first contribution in #55

Full Changelog: v0.3.6...v0.3.7

v0.3.6

What's Changed

• consider realm when activating project by name by @dlohmeier in #50
• remove PandaHubClient by @jthurner in #52
• project housekeeping by @jthurner in #51
• access of correct collection in get_net_ids by @dlohmeier in #54
• Fix upgrade project version: early exit, log spam by @jthurner in #57
• bump version to 0.3.6 by @jthurner in #58

Full Changelog: v0.3.5...v0.3.6

v0.3.5

What's Changed

• added collection_name to GetTimeseriesModel for API calls by @dlohmeier in #38

Full Changelog: v0.3.3...v0.3.5

pandahub v0.3.4

Changelog

• repaired tests
• added the possibility to add informations to a project
• added switches for CORS
• added more env parsing
• repaired the docker build
• added a pyproject.toml

pandahub v0.3.3

What's Changed

• Added registry class to JSON Decoder upon calling json.loads by @dlohmeier in #19
• Additional project entries by @dlohmeier in #43
• alleviate race condition creating non-unique net-ids by @jthurner in #45
• pandas 2.0 compatibility for timeseries compression by @julffers in #42
• make variants more robust by @jthurner in #46
• add mongodb_indexes and datatypes as arguments to PandaHub() by @jthurner in #48

Full Changelog: v0.3.2...v0.3.3

pandahub v0.3.2

• add support for timeseries collections
• fix: activate created project by id

Full Changelog: v0.3.1...v0.3.2

pandahub v0.3.1

• fix: filter asymmetric loads/sgens when getting subnets
• create indexes with project by default
  Full Changelog: v0.3.0...v0.3.1

pandahub v0.3.0

What's Changed

• Update fastapi-stack by @jthurner in #32

Full Changelog: v0.2.10...v0.3.0

BREAKING CHANGES:

• existing user collections have to be migrated (recreate with id field moved into _id - see below)
• changed variable names in fastapi-mail: MAIL_TLS -> MAIL_STARTTLS and MAIL_SSL -> MAIL_SSL_TLS
• applications integrating pandahub must adapt to changes introduced since fastapi-users v9 and fastapi v0.75

User database migration

pandahub v0.3.0 introduces a breaking change in the way user documents are stored. Log-ins will throw a backend error if you run pandahub >= v0.3.0 against a database containing users created with previous versions.

To migrate an existing database:

    from pandahub import PandaHub
    from pandahub.lib.database_toolbox import migrate_userdb_to_beanie
    ph = PandaHub()
    migrate_userdb_to_beanie(ph)

This migration breaks logins for pandahub < 0.3.0. Trying to create new users with an old pandahub version on a migrated database will fail if at least one user already exists. A backup of the original user collection will be created as users_fa9_%Y-%m-%d_%H-%M.

Background

fastapi-users, which pandahub depends on for user management, introduces a couple of breaking changes in version 10. With the switch to Beanie ODM for mongodb support, user ids are now saved on the _id field instead of id:

fastapi v9

{'_id': ObjectId('637f2341304330c7511e4056'), 'id': UUID('2b52cf59-f5c3-433c-963c-834fbe9edbeb'), 'email': 'my@user.com'}

fastapi v10+

{'_id': UUID('2b52cf59-f5c3-433c-963c-834fbe9edbeb'), 'email': 'my@user.com'}

To migrate, drop the index on id and move the id field into _id:

db["user_management"]["users"].drop_index("id_1")
db["user_management"]["users"].aggregate([{'$addFields': {'_id': '$id'}}, {'$unset': 'id'}, {'$out': 'users'}]

To roll back to pre-0.3.0, apply the reverse transformation:

db["user_management"]["users"].aggregate([{'$addFields': {'id': '$_id'}}, {'$unset': '_id'}, {'$out': 'users'}]`