Use a single common data source for public transports stops

Now makes use of Navitia opendata dumps to cover France. Fixes #65.

README.md

@@ -74,8 +74,7 @@ which covers Paris. If you want to run the script using some other location,
 you might have to change these files by matching datasets.
 
 * [LaPoste Hexasmal](https://datanova.legroupe.laposte.fr/explore/dataset/laposte_hexasmal/?disjunctive.code_commune_insee&disjunctive.nom_de_la_commune&disjunctive.code_postal&disjunctive.libell_d_acheminement&disjunctive.ligne_5) for the list of cities and postal codes in France.
-* [RATP (Paris) stations](https://data.ratp.fr/explore/dataset/positions-geographiques-des-stations-du-reseau-ratp/table/?disjunctive.stop_name&disjunctive.code_postal&disjunctive.departement) for the list of subway/tram/bus stations with their positions in Paris and nearby areas.
-* [Tcl (Lyon) stations](https://download.data.grandlyon.com/wfs/rdata?SERVICE=WFS&VERSION=2.0.0&outputformat=GEOJSON&maxfeatures=4601&request=GetFeature&typename=tcl_sytral.tclarret&SRSNAME=urn:ogc:def:crs:EPSG::4326) for the list of subway/tram/bus stations with their positions in Paris and nearby areas.
+* [Navitia public transport datasets](https://navitia.opendatasoft.com/explore/?sort=modified&refine.geographicarea=France) for the list of subway/tram/bus stations with their positions in France. These are the `stops_fr-*.txt` files, extracted from the `NTFS` datasets for each region.
 
 Both datasets are licensed under the Open Data Commons Open Database License
 (ODbL): https://opendatacommons.org/licenses/odbl/.

flatisfy/__main__.py

@@ -164,7 +164,7 @@ def main():
         if args.cmd == "build-data":
             force = True
 
-        data.preprocess_data(config, force=force)
+        if data.preprocess_data(config, force=force):
             LOGGER.info("Done building data!")
 
         if args.cmd == "build-data":

flatisfy/data.py

@@ -43,6 +43,7 @@ def preprocess_data(config, force=False):
 
     :params config: A config dictionary.
     :params force: Whether to force rebuild or not.
+    :return bool: Whether data have been built or not.
     """
     # Check if a build is required
     get_session = database.init_db(config["database"], config["search_index"])
@@ -53,7 +54,7 @@ def preprocess_data(config, force=False):
         )
         if is_built and not force:
             # No need to rebuild the database, skip
-            return
+            return False
         # Otherwise, purge all existing data
         session.query(PublicTransport).delete()
         session.query(PostalCode).delete()
@@ -67,6 +68,7 @@ def preprocess_data(config, force=False):
             )
         with get_session() as session:
             session.add_all(data_objects)
+    return True
 
 
 @hash_dict
@@ -88,7 +90,7 @@ def load_data(model, constraint, config):
         areas = []
         # Get areas to fetch from, using postal codes
         for postal_code in constraint["postal_codes"]:
-            areas.append(data_files.french_postal_codes_to_iso_3166(postal_code))
+            areas.append(data_files.french_postal_codes_to_quarter(postal_code))
         # Load data for each area
         areas = list(set(areas))
         for area in areas:

flatisfy/data_files/__init__.py

@@ -3,10 +3,14 @@
 Preprocessing functions to convert input opendata files into SQLAlchemy objects
 ready to be stored in the database.
 """
+import csv
+import io
 import json
 import logging
 import os
 
+from backports import csv
+
 from flatisfy.models.postal_code import PostalCode
 from flatisfy.models.public_transport import PublicTransport
 
@@ -15,18 +19,20 @@ LOGGER = logging.getLogger(__name__)
 MODULE_DIR = os.path.dirname(os.path.realpath(__file__))
 
 
-def french_postal_codes_to_iso_3166(postal_code):
+def french_postal_codes_to_quarter(postal_code):
     """
-    Convert a French postal code to the main subdivision in French this postal
-    code belongs to (ISO 3166-2 code).
+    Convert a French postal code to the main quarter in France this postal
+    code belongs to.
 
     :param postal_code: The postal code to convert.
-    :returns: The ISO 3166-2 code of the subdivision or ``None``.
+    :returns: The quarter of France or ``None``.
     """
+    departement = postal_code[:2]
+
     # Mapping between areas (main subdivisions in French, ISO 3166-2) and
     # French departements
     # Taken from Wikipedia data.
-    area_to_departement = {
+    department_to_subdivision = {
         "FR-ARA": ["01", "03", "07", "15", "26", "38", "42", "43", "63", "69",
                    "73", "74"],
         "FR-BFC": ["21", "25", "39", "58", "70", "71", "89", "90"],
@@ -44,17 +50,30 @@ def french_postal_codes_to_iso_3166(postal_code):
         "FR-PDL": ["44", "49", "53", "72", "85"],
         "FR-PAC": ["04", "05", "06", "13", "83", "84"]
     }
+    subdivision_to_quarters = {
+        'FR-IDF': ['FR-IDF'],
+        'FR-NW': ['FR-BRE', 'FR-CVL', 'FR-NOR', 'FR-PDL'],
+        'FR-NE': ['FR-BFC', 'FR-GES', 'FR-HDF'],
+        'FR-SE': ['FR-ARA', 'FR-COR', 'FR-PAC', 'FR-OCC'],
+        'FR-SW': ['FR-NAQ']
+    }
 
-    departement = postal_code[:2]
-    return next(
+    subdivision = next(
         (
             i
-            for i in area_to_departement
-            if departement in area_to_departement[i]
+            for i, departments in department_to_subdivision.items()
+            if departement in departments
+        ),
+        None
+    )
+    return next(
+        (
+            i
+            for i, subdivisions in subdivision_to_quarters.items()
+            if subdivision in subdivisions
         ),
         None
     )
-
 
 
 def _preprocess_laposte():
@@ -69,8 +88,8 @@ def _preprocess_laposte():
     raw_laposte_data = []
     # Load opendata file
     try:
-        with open(
-            os.path.join(MODULE_DIR, data_file), "r"
+        with io.open(
+            os.path.join(MODULE_DIR, data_file), "r", encoding='utf-8'
         ) as fh:
             raw_laposte_data = json.load(fh)
     except (IOError, ValueError):
@@ -82,7 +101,7 @@ def _preprocess_laposte():
     for item in raw_laposte_data:
         fields = item["fields"]
         try:
-            area = french_postal_codes_to_iso_3166(fields["code_postal"])
+            area = french_postal_codes_to_quarter(fields["code_postal"])
             if area is None:
                 LOGGER.info(
                     "No matching area found for postal code %s, skipping it.",
@@ -104,72 +123,45 @@ def _preprocess_laposte():
     return postal_codes_data
 
 
-def _preprocess_ratp():
+def _preprocess_public_transport():
     """
-    Build SQLAlchemy objects from the RATP data (public transport in Paris,
-    France).
+    Build SQLAlchemy objects from the Navitia public transport data.
 
     :return: A list of ``PublicTransport`` objects to be inserted in database.
     """
-    data_file = "ratp.json"
-    LOGGER.info("Building from %s data.", data_file)
+    DATA_FILES = {
+        "FR-IDF": "stops_fr-idf.txt",
+        "FR-NW": "stops_fr-nw.txt",
+        "FR-NE": "stops_fr-ne.txt",
+        "FR-SW": "stops_fr-sw.txt",
+        "FR-SE": "stops_fr-se.txt"
+    }
 
-    ratp_data_raw = []
+    public_transport_data = []
     # Load opendata file
+    for area, data_file in DATA_FILES.items():
+        LOGGER.info("Building from public transport data %s." % data_file)
         try:
-        with open(os.path.join(MODULE_DIR, data_file), "r") as fh:
-            ratp_data_raw = json.load(fh)
-    except (IOError, ValueError):
-        LOGGER.error("Invalid raw RATP opendata file.")
+            with io.open(os.path.join(MODULE_DIR, data_file), "r",
+                         encoding='utf-8') as fh:
+                filereader = csv.reader(fh)
+                next(filereader, None)  # Skip first row (headers)
+                for row in filereader:
+                    public_transport_data.append(PublicTransport(
+                        name=row[2],
+                        area=area,
+                        lat=row[3],
+                        lng=row[4]
+                    ))
+        except (IOError, IndexError):
+            LOGGER.error("Invalid raw opendata file: %s." % data_file)
             return []
 
-    # Process it
-    ratp_data = []
-    for item in ratp_data_raw:
-        fields = item["fields"]
-        ratp_data.append(PublicTransport(
-            name=fields["stop_name"],
-            area="FR-IDF",
-            lat=fields["coord"][0],
-            lng=fields["coord"][1]
-        ))
-    return ratp_data
-
-
-def _preprocess_tcl():
-    """
-    Build SQLAlchemy objects from the Tcl data (public transport in Lyon,
-    France).
-
-    :return: A list of ``PublicTransport`` objects to be inserted in database.
-    """
-    data_file = "tcl.json"
-    LOGGER.info("Building from %s data.", data_file)
-
-    tcl_data_raw = []
-    # Load opendata file
-    try:
-        with open(os.path.join(MODULE_DIR, data_file), "r") as fh:
-            tcl_data_raw = json.load(fh)
-    except (IOError, ValueError):
-        LOGGER.error("Invalid raw Tcl opendata file.")
-        return []
-
-    # Process it
-    tcl_data = []
-    for item in tcl_data_raw["features"]:
-        tcl_data.append(PublicTransport(
-            name=item["properties"]["nom"],
-            area="FR-ARA",
-            lat=item["geometry"]["coordinates"][1],
-            lng=item["geometry"]["coordinates"][0]
-        ))
-    return tcl_data
+    return public_transport_data
 
 
 # List of all the available preprocessing functions. Order can be important.
 PREPROCESSING_FUNCTIONS = [
     _preprocess_laposte,
-    _preprocess_ratp,
-    _preprocess_tcl
+    _preprocess_public_transport
 ]

requirements.txt

@@ -1,5 +1,6 @@
 appdirs
 arrow
+backports.csv
 bottle
 bottle-sqlalchemy
 canister