1. 14 May, 2020 1 commit
  2. 10 May, 2020 1 commit
    • Volker Krause's avatar
      Change the way we store timezones in the static database · 14800de9
      Volker Krause authored
      So far this were offsets into the the IANA string table, now it's a flat
      enum. The old way needed 13 bits per record, the new only needs 9 bit, at
      the cost of an extra ~800 bytes for an offset table to get back to IANA
      names. This however quickly pays of when storing large quantities, which
      we do (~37k in the current database, more in the upcoming experiments for
      an efficient geo coordinate to timezone mapping).
      14800de9
  3. 27 Mar, 2019 2 commits
  4. 16 Mar, 2019 1 commit
  5. 13 Mar, 2019 2 commits
  6. 26 Feb, 2019 1 commit
  7. 23 Feb, 2019 1 commit
  8. 22 Feb, 2019 1 commit
  9. 27 Jan, 2019 1 commit
  10. 26 Jan, 2019 1 commit
  11. 27 Dec, 2018 1 commit
  12. 17 Nov, 2018 2 commits
  13. 08 Sep, 2018 1 commit
    • Volker Krause's avatar
      Improve string normalization, and use that for airport name matching · 9fa1a980
      Volker Krause authored
      So far we were just doing case folding, now we also do Unicode
      decomposition to remove diacritic marks. This reduces the airport
      string table size by ~5% without compromising quality.
      
      This approach should also be helping for matching non-ASCII names in
      IATA boarding passes to their normal spelling.
      9fa1a980
  14. 26 May, 2018 1 commit
  15. 21 May, 2018 1 commit
  16. 12 May, 2018 3 commits
  17. 04 Apr, 2018 1 commit
  18. 17 Mar, 2018 1 commit
  19. 16 Jan, 2018 1 commit
  20. 13 Jan, 2018 1 commit
    • Volker Krause's avatar
      Make the airport lookup a bit more clever · 436d6e24
      Volker Krause authored
      The previous method of only looking of unique fragments of the name worked
      reasonably well in Europe, but not in the US. In the US there are many
      (regional) airports with overlapping name fragments so that even major
      international airports could not be looked up properly. We now also
      consider non-unique fragments if they result in a unique subset. This
      doesn't impact the database size much, but it significantly improves
      the detection of US airports.
      
      This also re-generates the database from latest Wikidata data, reducing
      the IATA code conflicts from ~250 to 140 due to upstream fixes.
      436d6e24
  21. 10 Dec, 2017 2 commits