Change Log

For the most up-to-date version of Specklia, please consult PyPI.

v1.8.57

  • Provide a more helpful message to the end user when the search polygon is too large.

v1.8.48-56

  • Routine dependency upgrades.

v1.8.47

  • Added deduplication tools.

v1.8.46

  • Tweaks to the ingestion of CryoTEMPO data based on experience gained while ingesting Baseline B products.

v1.8.45

  • Routine dependency upgrades.

v1.8.44

  • Fixed issues uncovered during the ingestion of CryoTEMPO POCA Baseline B products.

v1.8.40-43

  • Routine dependency upgrades.

v1.8.39

v1.8.37-38

  • Routine dependency upgrades.

v1.8.36

  • Tweaked user activity tracking database configuration.

v1.8.32-35

  • Routine dependency upgrades.

v1.8.30-31

  • Fixed a mistake in the Austfonna example notebook.

v1.8.29

  • Ingested the RGI v7 1st and 2nd order region polygons.

v1.8.28

  • Added two new, more detailed examples of the use of Specklia to the documentation.

v1.8.25-27

  • Routine dependency upgrades.

v1.8.24

  • Slightly improved the resilience of websocket connections.

v1.8.22-23

  • Routine dependency upgrades.

v1.8.21

  • Fixed an issue where an Unhandled Exception would be reported if multiple datasets were ingested in the same API call, but only some of them had the required columns.

  • Fixed some misleading statements in the documentation regarding ingest.

v1.8.20

  • Minor updates to ingestion for a polygon datasets.

v1.8.19

  • Excessively long requests are now truncated in the serverside logs.

v1.8.18

  • Fixed an issue where geospatial indexes were being incorrectly created in OLTP datasets.

v1.8.17

  • Accelerated ingest by updating metadata in batches.

v1.8.15-16

  • Routine dependency upgrades.

v1.8.14

  • client.utilities.save_gdf_as_tiff is now more robust in cases where gridded datasets are sparse.

v1.8.13

  • Fixed an issue where Geospatial indexes were being incorrectly created for OLTP datasets.

v1.8.12

  • Fixed an issue which unnecessarily halved the speed of OLTP dataset queries.

v1.8.11

  • Removed some indexing that proved not to be cost-effective.

v1.8.10

  • Sources are now correctly deleted when datasets are deleted.

v1.8.9

  • Significantly improved logic around the updating of the metadata returned by list_datasets(). Datasets created before this point may have innaccurate values for their min and max custom column values.

v1.8.8

  • added /usage route, which allows retrieval of usage information for a particular group.

v1.8.7

  • Routine dependency upgrade.

v1.8.6

  • Fixed an issue where float valued custom columns were not being properly typecast on metdata update.

v1.8.5

  • Routine dependency upgrade.

v1.8.4

  • Compatibility fixes for cs2eo.org integration.

v1.8.3

  • Fixed an issue relating to integer handling in MongoDB.

v1.8.2

  • Routine dependency upgrade.

v1.8.1

  • Bundle of MongoDB compatibility fixes.

v1.8.0

  • Specklia Datasets can now be stored using one of two storage technologies: Online Analysis Processing (OLAP) and Online Transaction Processing (OLTP). The storage technology is selected when Datasets are created, and cannot be changed afterwards. OLAP datasets have cheaper storage costs and no storage limit, but are slower to query. OLTP datasets are more expensive and have a storage limit, but are much faster to query.

v1.7.41

  • Additional handling for antimeridian issues.

v1.7.37-40

  • Routine dependency upgrades.

v1.7.35-36

  • Database optimisations.

v1.7.34

  • Specklia will now return a client-visible error when a query that is too large for the server’s memory is submitted.

v1.7.33

  • Re-enabled documentation generation after a breaking change.

v1.7.32

  • Fixes related to MongoDB transition.

v1.7.31

  • Four internal tables transitioned to MongoDB.

v1.7.26-30

  • Routine dependency upgrades.

v1.7.25

  • Initial integration of MongoDB.

v1.7.23-24

  • Routine dependency upgrades.

v1.7.22

  • Documentation tweaks.

v1.7.21

  • Specklia now contains the Randolph Glacier Inventory v7.0 and IMBIE Basin polygon datasets.

v1.7.19-20

  • Routine dependency upgrades.

v1.7.18

  • Fixed some documentation generation issues.

v1.7.17

  • Routine dependency upgrade.

v1.7.16

v1.7.8-15

  • Routine dependency upgrades.

v1.7.7

  • Maximum number of points per ingest increased to 10M.

v1.7.5-7

  • Routine dependency upgrades.

v1.7.4

  • Newly ingested data is now no longer checked for duplicate sources, due to cost constraints.

v1.7.2-3

  • Significantly reduced costs associated with the querying of the sources table.

v1.7.1

  • Fixed containerisation and CICD issues.

v1.7.0

  • When ingesting, Specklia now only checks for duplicate sources that overlap the newly ingested data. Previously, the full dataset was checked for duplicate sources.

  • Minor version bump to rapidly resolve CICD problem.

v1.6.23

  • Routine dependency upgrade.

v1.6.22

  • Re-enabled documentation after websocket transport change.

v1.6.21

  • Fixed an issue where websocket transmission of very small messages would sometimes fail.

v1.6.20

  • Websocket transport changed again to enable API to be directly used within common web browsers.

  • Minor memory optimisation.

v1.6.11-19

  • Routine dependency upgrades.

v1.6.10

  • Re-enabled documentation generation after breaking API change.

v1.6.9

  • Websocket transport change to enable API to be used directly by Javascript.

v1.6.8

  • Routine dependency upgrades.

v1.6.7

  • Fixed an issue where points could not be added to integer typed columns.

v1.6.4-6

  • Routine dependency upgrades.

v1.6.3

  • Further increased query speed.

  • Columns other than Polygons are now properly type-checked on ingest.

  • Fixed a bug that arises when attempting to add data to a dataset that only contains polygon-typed custom columns.

  • Fixed a misleading error message related to polygon ingest.

v1.6.1-2

  • Routine dependency upgrades.

v1.6.0

  • Minor version bump to resolve CI/CD issues.

v1.5.38

  • It is now possible to query datasets using multipolygons.

  • When querying a dataset that does not exist, users will now receive the correct error message.

  • Error messaging will now explain polygon invalidity when a polygon is invalid.

  • Improved error reporting when user does not have permission to add points to a dataset.

  • Swapped out Esri imagery source for Geoportail France imagery source in documentation examples.

  • Fixed an incorrect query typehint in the client.

v1.5.35-1.5.37

  • Routine dependency upgrades.

v1.5.34

  • Fixed an issue where documentation was not being generated, and made some client.query_dataset arguments optional.

v1.5.33

  • It is now possible to retrieve only the source information for a dataset using the source_information_only flag within client.query_dataset.

v1.5.32

v1.5.31

  • Sources now record the point count, spatial and temporal extent of the source.

v1.5.30

  • Removed row_id from specklia datasets, significantly reducing data volume.

v1.5.29

  • Routine dependency upgrades.

v1.5.28

  • It is now possible to retrieve polygons from Specklia using polygon- and multipolygon-valued custom columns.

v1.5.27

  • Routine dependency upgrades.

v1.5.26

  • It is now possible to create a polygon- and multipolygon-valued custom column.

v1.5.25

  • Slightly accelerated querying.

v1.5.16-24

  • Routine dependency upgrades.

v1.5.16

  • Simplified and accelerated backend.

v1.5.13-15

  • Routine dependency upgrades & CI/CD fixes.

v1.5.12

  • Added a tool to the client to allow the conversion of raster data stored within Specklia back to GeoTIFFs.

v1.5.10-11

  • Routine dependency upgrades.

v1.5.10

  • Handled an additional rarely occuring timeout.

v1.5.3-9

  • Routine dependency upgrades.

v1.5.2

  • Handled an additional rarely occuring timeout.

v1.5.1

  • Correctly handled a 3rd party originated ingest error.

v1.5.0

  • New version to reset CI/CD chain problems.

v1.4.6-35

  • Routine dependency upgrades.

v1.4.6

  • Reduced Specklia query time, in particular for small queries.

  • Fitted logging for detailed timing of the query process.

v1.4.5

  • Fixed a documentation issue where EPSG4326 Dataset Coverage was being plotted incorrectly around the South Pole.

v1.4.1-4

  • Routine dependency upgrades.

v1.4.0

  • Significant internal refactor.

v1.3.11

  • Reduced running costs.

  • Tweaked timeouts.

  • Prevented the ingest of data of length zero.

  • Prevented the ingest of data containing exclusively NaN or Inf values within a column.

v1.3.9-10

  • Routine dependency upgrades.

v1.3.8

  • Fixed error in timestamp calculation during ingestion.

v1.3.7

  • Significantly increased the amount of RAM available to an ingest worker.

  • Added a cap on the number of points that the client can request to have ingested at once.

v1.3.4-6

  • Routine dependency upgrades.

v1.3.3

  • Updated Examples within documentation.

v1.3.2

  • Slightly accelerated spatial indexing.

  • Accelerated data transfer within Specklia.

  • Fixed a rare spatial indexing error.

  • Suppressed some errors related to early but benign websocket closure.

  • Fixed some minor documentation typos.

v1.3.1

  • Routine dependency upgrade.

v1.3.0

  • Minor version change to recognise that Antarctic ingests now work correctly, so data can be ingested worldwide.

  • Capped the amount of queries each user can submit within a rolling 30 day window.

v1.2.40

  • Fixed an issue where the coverage polygon for Antarctica was too large to add to the metadata table.

  • Fixed an issue where large-valued s2 cell IDs were being incorrectly calculated.

  • Fixed an issue where newly created datasets would not have the expected foreign key relationships.

v1.2.39

  • Reduced data storage cost.

  • Made it impossible to create a group with the same name as the ‘all_users’ group using the API.

v1.2.38

  • Reversed a dask dependency upgrade to address an issue with map_partitions.

v1.2.35-37

  • Routine dependency upgrades.

v1.2.34

  • Fixed an issue that prevented ingestion of datasets that produced coverage polygons including the north or south poles.

v1.2.20-33

  • Routine dependency upgrades.

v1.2.20

  • Accelerated evaluation of Example Notebooks and removed unecessary NaN handling.

v1.2.20

  • Fixed a race condition that occurred on massively parallel ingests.

v1.2.19

  • Fixed an issue where calculation of the min and max fields within dataset metadata would fail when the dataset contained both string and numeric custom columns.

v1.2.18

  • Improved schema error reporting for the /ingest and /query routes.

v1.2.17

  • Fixed an issue where the min and max fields within dataset metadata would be incorrectly calculated when performing parallel ingests.

v1.2.16

  • Improved schema error reporting for the /users and /groupmembership routes.

v1.2.15

  • Fixed CICD Error causing the use of the wrong JWT token key.

v1.2.14

  • The dataframe returned by list_datasets() now has more appropriate dtypes for epsg4326_coverage and various timestamp columns.

v1.2.11-13

  • Routine dependency upgrades.

v1.2.10

  • Improved schema error reporting for the /groups route.

v1.2.9

  • Fixed an issue that arose when trying to ingest a GeoDataFrame with a non-trivial index.

  • Fixed an issue where all_users admins could not list users in and delete groups that they were not explicitly a part of.

v1.2.8

  • Cloud hosting cost optimisations.

v1.2.7

  • Improved schema error reporting for the /metadata route.

v1.2.6

  • Added two new examples of Specklia use to the documentation.

v1.2.5

  • Client immediately attempts to connect to server when instantiated, ensuring the user is promptly informed if their API key is invalid.

  • Change log now visible in documentation.

v1.2.4

  • list_datasets() now returns meaningful strings for user privileges, instead of numbers.

v1.2.3

  • Routine dependency upgrade.

v1.2.2

  • It is no longer possible to cease to be an admin of your personal group.

v1.2.1

  • Fixed an issue where users attempting to obtain an API key would sometimes experience an Internal Server Error.

  • Introduced custom HTTP 500 and 404 response pages.

v1.2.0

  • Significantly improved documentation.

  • Notified ESA to commence beta testing.

v1.1.0

  • First version released to PyPI.

Versioning and releases

Specklia follows the common v<major>.<minor>.<patch> model of version numbering, together with Continuous Deployment. We will frequently release versions that change the <patch> version number in order to fix bugs. <minor> version number changes indicate changes to Specklia’s functionality, while <major> version changes indicate re-writes or other similarly disruptive changes.

As Specklia is currently in development, these Change Logs may be incomplete. If you have questions about a specific Specklia version, please email support@earthwave.co.uk