tommy/libpostal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[docs] Adding descriptions of some of the new dictionary types

Browse Source
This commit is contained in:
Al
2016-03-27 20:22:50 -04:00
parent 8a105ae339
commit 37f6e6034d
1 changed files with 16 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
resources/dictionaries/README.md
View File

@@ -22,6 +22,7 @@ Each language can define one or more dictionaries (sometimes called "gazetteers"
- **ambiguous_expansions.txt**: e.g. "E" could be expanded to "East" but could
be "E Street", so if the string is encountered, it can either be left alone or expanded. In general, single-letter abbreviations in most languages should also be added to ambiguous_expansions.txt since single letters are also often initials
- **building_types.txt**: strings indicating a building/house
- **categories.txt**: category strings e.g. from [Nominatim Special Phrases](http://wiki.openstreetmap.org/wiki/Nominatim/Special_Phrases) expected to be used in searches like "restaurants in Manhattan". Singular and plural forms can be included here.
- **company_types.txt**: company suffixes like "Inc" or "GmbH"
- **concatenated_prefixes_separable.txt**: things like "Hinter..." which can
be written either concatenated or as separate tokens
@@ -33,16 +34,22 @@ say, part of a street name
suffix can be either concatenated to the main token or separated
- **directionals.txt**: strings indicating directions (cardinal and
lower/central/upper, etc.)
- **level_types.txt**: strings indicating a particular floor
- **no_number.txt**: strings like "no fixed address"
- **nulls.txt**: strings meaning "not applicable"
- **personal_suffixes.txt**: post-nominal suffixes, usually generational
like Jr/Sr
- **personal_titles.txt**: civilian, royal and military titles
- **entrance.txt**: string indicating an entrance, usually just the word "entrance" and its appropriate abbreviations.
- **house_number.txt**: strings that may be added as part of the house/building number (for languages like Spanish where it's common to say "No. 123" or "No. Ext. 123" for the house/building number instead of just "123" as in English).
- **level_types_basement.txt**: strings indicating a basement level.
- **level_types_mezzanine.txt**: strings indicating a mezzanine level.
- **level_types_numbered.txt**: strings indicating a numbered level of a building (numbered).
- **level_types_standalone.txt**: strings indicating a level/floor of a building that can stand on their own without a number like "ground floor", etc.
- **level_types_sub_basement.txt**: strings indicating a sub-basement level.
- **no_number.txt**: strings like "sin número" used for houses with no number.
- **nulls.txt**: strings meaning "not applicable" e.g. in spreadsheets or database fields that might have missing values
- **personal_suffixes.txt**: post-nominal suffixes, usually generational e.g. Junior/Senior in English or der Jungere in German.
- **personal_titles.txt**: civilian, royal, clerical, and military titles e.g. "Saint", "General", etc.
- **place_names.txt**: strings found in names of places e.g. "theatre",
"aquarium", "restaurant". [Nominatim Special Phrases](http://wiki.openstreetmap.org/wiki/Nominatim/Special_Phrases) is a great resource for this.
- **post_office.txt**: strings like "p.o. box"
- **qualifiers.txt**: strings like "township"
- **staircase.txt**: strings indicating a staircase, usually just the word "staircase" or "stair".
- **stopwords.txt**: prepositions and articles mostly, very common words
which may be ignored in some contexts
- **street_types.txt**: words like "street", "road", "drive" which indicate
@@ -53,7 +60,9 @@ just be treated as string replacement.
- **toponyms.txt**: abbreviations for certain abbreviations relating to
toponyms like regions, places, etc. Note: GeoNames covers most of these.
In most cases better to leave these alone
- **unit_types.txt**: strings indicating an apartment or unit number
- **unit_directions.txt**: phrases to indicate which side of the building the apartment/unit is on, usually along the lines of "left", "right", "front", "rear".
- **unit_types_numbered.txt**: strings indicating a apartment or unit e.g. we expect a number to follow (or in some languages, precede) strings like "flat", "apt", "unit", etc.
- **unit_types_standalone.txt**: for unit type that can stand on their own without an accompanying number e.g. "penthouse".
Most of the dictionaries have been derived using the following process: