Released 2019-06-04
- Fixed an error when the query is pure regexp and it did not match anything.
Released 2018-05-30
- Refactored
conflate.pyinto seven smaller files. - Added a simple kd-tree based geocoder for countries and regions. Controlled by the
regionsparameter in a profile. - You can filter by regions using
-rargument or"regions"list in an audit file. - Using the new
nwrquery type of Overpass API. - Reduced default
max_request_boxesto four. - New argument
--alt-overpassto use Kumi Systems' server (since the main one is blocked in Russia). - Better handling of server runtime errors.
- Find matches in OSM with
--list <result.csv>. - Control number of nearest points to check for matches with
nearest_pointsprofile parameter. - When you have dataset ID in an URL or other tag, use
find_refprofile function to match on it.
Released 2018-04-26
- Fixed processing of
''tag value. - More that 3 duplicate points in a single place are processed correctly.
- Now you can
yieldpoints from a profile instead of making a list. - Not marking nodes with
movein the audit file as modified, unless we move them.
Released 2018-04-19
- Fixed bug in categories building.
- Fixed threshold for tags in duplicates check.
- Now the script prints "Done" when finished, to better measure time.
Released 2018-03-20
- "Similar tags" now means at least 66% instead of 50%.
- Instead of removing all duplicates, conflating them and removing only unmatched.
Released 2018-03-15
-
Support for categories:
category_tagandcategoriesparameters in a profile. -
LibOsmium-based C++ filtering script for categories.
-
More than one tag value works as "one of":
[('amenity', 'cafe', 'restaurant')]. -
Query can be a list of queries, providing for "OR" clause. An example:
[[('amenity', 'swimming_pool')], [('leisure', 'swimming_pool')]] -
Parameters for profiles, using
-pargument. -
No more default imports solely for profiles, import
zipfileyouself now. -
Remarks for source points, thanks @nixi.
-
Better error message for Overpass API timeouts.
-
Lifecycle prefixes are conflated, e.g.
amenity=*andwas:amenity=*. -
Dataset is checked for duplicates, which are reported (see
-d) and removed. -
Support GeoJSON input (put identifiers into
idproperty).
Released 2017-12-29
- Fix error in applying audit json after conflating
contact:namespace.
Released 2017-12-27
- Addr:full tag is not set when addr:housenumber is present.
- Whitespace is stripped from tag values in a dataset.
- Conflate
contact:namespace.
Released 2017-12-20
- Support force creating points with
audit['create']. - Fix green colour for created points in JSON.
- Make
--outputoptional and remove the default.
Released 2017-11-23
- Checking moveability for json output (
-m) for cf_audit. - Support for cf_audit json (
-a).
Released 2017-10-06
- Use
-vfor debug messages and-qto suppress informational messages. - You can run
conflate/conflate.pyas a script, again. - Profiles: added "override" dict with dataset id → OSM POI name or id like 'n12345'.
- Profiles: added "matched" function that returns
Falseif an OSM point should not be matched to dataset point (fixes #6). - Profiles:
master_tagsis no longer mandatory. - If no
master_tagsspecified in a profile, all tags are now considered non-master. - When a tag value was
None, the tag was deleted on object modification. That should be done only on retagging non-matched objects. - OSM objects filtering failed when a query was a string.
Released 2017-06-07
The initial PyPi release with all the features.