TODO

Todo list for the OpenTrep project
----------------------------------

See now http://github.com/trep/opentrep/issues/

* [25/06/2010] For date parsing, RememberTheMilk (RTM) gathers on a page
  all the formats supported by its parser:
  http://www.rememberthemilk.com/help/answers/basics/dateformat.rtm

* [06/09/2009] Add the data type (e.g., "GEO", "TIM", "AIR", "PAX")
  in front of the Xapian document data (when indexing).

* [06/09/2009] Create a new table for time-related data (e.g.,
  "tue" for the code, and "en, tuesday" or "fr, mardi" for
  the names).
  Write the C++ code to take that new data into account.

* DONE / NA
  [06/06/2009] Alter the parser so that when a suggestion is made for
  a string, and when the suggestion for the same string minus a word
  is different, the Levenshtein distance between the original suppresed
  word and the "suggested" one (defined as the difference between the
  suggestions for the two strings) is something like nb_of_letters/3.

* DONE
  [06/09/2009] To allow searches like "ru" or "france", just add the
  corresponding (country) keywords to the main airport of that country
  (e.g., "par" for "fr"/"france" and "lon" for "uk"/"gb").

* DONE
  [06/09/2009] Create a table with the states and countries.
  Write the C++ code adding all the state- and country-related keywords
  to the geogrphical index (e.g., add the "texas" keyword to the "paris, tx"
  element).

* [01/08/2009] Write a (Python-based) PSP page, in order to test the
  different locales of the browsers.
  The Python (PSP) page has been created, but there is still some work
  to do in order to adapt it to the new API (with extra and alternate
  locations).

* [14/08/2009] With the ICU library, check the encoding of the input,
  and convert it to Unicode if needed (see the test/i18n/icuustring
  and test/i18n/icuconv samples for example). First detect and convert
  hard-coded strings, then do it on the output of PSP pages.

* DONE
  [14/08/2009] Write a transliterator, taking UTF-8 Cyrillic input
  (e.g., Russian and/or Ukrainian) and romanising/transliterating
  it. Note that there is good transliterator sample within the ICU
  Subversion repository
  (http://source.icu-project.org/repos/icu/icu/trunk/source/samples/translit).

* DONE
  [14/08/2009] With the ICU library, convert UTF-8 encoded strings to
  lower cases.

* DONE
  [01/08/2009] Finish the work on bringing extra and additional
  Location objects into the API.

* DONE
  [01/08/2009] In the search batch, when looking in the database, do
  it based on the airport/city code, rather than on the Xapian
  document ID. That way, no database update will be necessary when
  re-indexing, and any search from any Xapian index will find the
  corresponding result details within the database. The easiest way is
  to extract the first three letters of the Xapian document data.