jageocoder 2.2.0__tar.gz → 2.2.1__tar.gz

{jageocoder-2.2.0 → jageocoder-2.2.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: jageocoder
-Version: 2.2.0
+Version: 2.2.1
 Summary: A Japanese-address geocoder for Python.
 License: MIT
 Keywords: geocoder,Japanese address
@@ -17,6 +17,7 @@ Provides-Extra: doc
 Provides-Extra: test
 Requires-Dist: deprecated (>=1.2.13)
 Requires-Dist: docopt-ng
+Requires-Dist: flake8 ; extra == "dev"
 Requires-Dist: geographiclib (>=2.0)
 Requires-Dist: jaconv (>=0.3.9)
 Requires-Dist: marisa-trie (>=1.3.1)
@@ -24,6 +25,8 @@ Requires-Dist: pytest (>=8.4.1) ; extra == "test"
 Requires-Dist: requests
 Requires-Dist: rtree (>=1.0.0)
 Requires-Dist: sphinx-rtd-theme (>=1.2.0) ; extra == "doc"
+Requires-Dist: tox (>=4.0.0) ; extra == "dev"
+Requires-Dist: tox-uv ; extra == "dev"
 Requires-Dist: tqdm (>=4.0.0)
 Requires-Dist: twine (>=5.1.1) ; extra == "dev"
 Project-URL: Documentation, https://jageocoder.readthedocs.io/ja/latest/

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/__init__.py

@@ -17,7 +17,7 @@ running the following steps.
     >>> import jageocoder
     >>> jageocoder.init()
     >>> jageocoder.searchNode('新宿区西新宿2-8-1')
-    [{"node": {"id": ..., "name": "8番", "name_index": "8.番", "x": 139.6..., "y": 35.6..., "level": 7, "priority": 3, "note": "", "parent_id": ..., "sibling_id": ...}, "matched": "新宿区西新宿2-8-"}]
+    [{"node": {"id": ..., "name": "8番", "name_index": "8.番", "x": 139.6..., "y": 35.6..., "level": 7, "priority": 3, "note": "", "parent_id": ..., "sibling_id": ...}, "matched": "新宿区西新宿2-8-"}]  # noqa: E501
 """
 
 from importlib.metadata import version

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/__main__.py

@@ -127,7 +127,11 @@ def main():
 
     elif args['reverse']:
         from jageocoder.address import AddressLevel
-        jageocoder.init(db_dir=args['--db-dir'], mode='r', url=args['--url'])
+        jageocoder.init(
+            db_dir=args['--db-dir'],
+            mode='r',
+            debug=args['--debug'],
+            url=args['--url'])
         print(json.dumps(
             jageocoder.reverse(
                 x=float(args['<longitude>']),

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/aza_master.py

@@ -3,7 +3,6 @@ import datetime
 import json
 from logging import getLogger
 from pathlib import Path
-from typing import Optional
 import re
 from typing import Dict, Union
 

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/dbm/base_table.py

@@ -397,7 +397,7 @@ class BaseTable(object):
 
         # Create empty table
         cur = self.get_conn().cursor()
-        cur.execute(f"DROP TABLE IF EXISTS records")
+        cur.execute("DROP TABLE IF EXISTS records")
         cur.execute(self._get_create_table_statement())
         self.commit()
 
@@ -493,7 +493,6 @@ class BaseTable(object):
         if from_pos < 0 or to_pos > self.count_records():
             raise ValueError("Out of range.")
 
-        limits = to_pos - from_pos
         conn = self.get_conn()
         conn.row_factory = sqlite3.Row
         cur = conn.cursor()
@@ -762,7 +761,7 @@ class BaseTable(object):
                 raise ValueError(f"Field '{k}' is not in this table.")
             elif k == self.pos_field:
                 raise ValueError(
-                    f"The value of the pos field can not modified.")
+                    "The value of the pos field can not modified.")
 
             fields.append(k)
             if isinstance(self.schema[k], (dict, list)):
@@ -830,7 +829,7 @@ class BaseTable(object):
                     raise ValueError(f"Field '{k}' is not in this table.")
                 elif k == pos_field:
                     raise ValueError(
-                        f"The value of the pos field can not modified.")
+                        "The value of the pos field can not modified.")
 
                 cols.append(k)
                 if isinstance(self.schema[k], (dict, list)):

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/module.py

@@ -4,15 +4,14 @@ import shutil
 from typing import Any, Dict, List, Optional, Union
 import urllib.request
 from urllib.error import URLError
+import zipfile
 
 import jageocoder
-from .dataset import Dataset
 from .exceptions import JageocoderError
 from .local import LocalTree
 from .tree import AddressTree, get_db_dir
 from .remote import RemoteTree
 from .result import Result
-from .rtree import Index
 
 _tree: Optional[AddressTree] = None  # The default AddressTree
 logger = logging.getLogger(__name__)
@@ -193,7 +192,6 @@ def get_module_tree() -> AddressTree:
     AddressTree
         The singleton object.
     """
-    global _tree
     if _tree is None:
         raise JageocoderError("Tree is not initialized")
 
@@ -268,10 +266,17 @@ def install_dictionary(
     # Unzip the archive
     os.makedirs(db_dir, exist_ok=True)
     shutil.rmtree(db_dir)
-    shutil.unpack_archive(
-        filename=str(path),
-        extract_dir=str(db_dir),
-    )
+    # shutil.unpack_archive(
+    #     filename=str(path),
+    #     extract_dir=str(db_dir),
+    # )
+    with zipfile.ZipFile(path, "r") as zipf:
+        zipinfo_list = zipf.infolist()
+        n = len(zipinfo_list)
+        for index, zipinfo in enumerate(zipinfo_list):
+            logger.info(f"[{index+1}/{n}] Extractiong '{zipinfo.filename}'")
+            zipf.extract(zipinfo, db_dir)
+
     for readme_fname in ("README.txt", "README.md",):
         readme_path = os.path.join(db_dir, readme_fname)
         if os.path.exists(readme_path):

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/node.py

@@ -122,7 +122,7 @@ class AddressNodeTable(AbstractTable):
             notes = []
             for attr in re.split(r'(?<!\\)/', note):
                 try:
-                    k, v = re.split(r'(?<!\\):', attr, 1)
+                    k, v = re.split(r'(?<!\\):', attr, maxsplit=1)
                 except ValueError:
                     k, v = '', attr
 
@@ -539,7 +539,7 @@ class AddressNode(object):
         notes = []
         for attr in re.split(r'(?<!\\)/', self.note):
             try:
-                k, v = re.split(r'(?<!\\):', attr, 1)
+                k, v = re.split(r'(?<!\\):', attr, maxsplit=1)
             except ValueError:
                 k, v = '', attr
 

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/remote.py

@@ -3,7 +3,7 @@ from collections import OrderedDict
 import json
 from logging import getLogger
 import os
-from typing import Any, Dict, List, NoReturn, Optional, Tuple, Union
+from typing import Any, Dict, List, NoReturn, Optional, Union
 import uuid
 
 import requests

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/rtree.py

@@ -356,10 +356,25 @@ class Index(object):
         index.Rtree
             Created rtree index.
         """
-        import time
         file_idx = index.Rtree(str(treepath))  # Filename must be passed as str
         node_table: AddressNodeTable = self._tree.address_nodes
 
+        def _insert_node_to_rtree(
+            id: int,
+            coordinates: Tuple[float, float, float, float],
+        ) -> None:
+            file_idx.insert(
+                id=id,
+                coordinates=coordinates,
+            )
+
+        def _get_limit_bdr(cx: float, cy: float, radius: float) -> Tuple[float, float, float, float]:
+            # cx, cy を中心とし、半径 radius (m)の円を内包する BDR の (x0, y0, x1, y1) を返す
+            import math
+            delta_lat = radius / 111320.0
+            delta_lon = radius / (111320.0 * math.cos(math.radians(cy)))
+            return (cx - delta_lon, cy - delta_lat, cx + delta_lon, cy + delta_lat)
+
         max_id = AddressNode.ROOT_NODE_ID + node_table.count_records()
         registered_coordinates = set()
 
@@ -367,47 +382,49 @@ class Index(object):
         id = AddressNode.ROOT_NODE_ID
         with tqdm(total=max_id, mininterval=0.5, ascii=True) as pbar:
             mode = ""
+            bdr = None
+            first_block_node = None
+            parent_node = None
             sibling_id = AddressNode.ROOT_NODE_ID
             for node in node_table.get_nodes_by_id(
                     AddressNode.ROOT_NODE_ID, max_id):
                 id = node.id
                 pbar.update(1)
 
-                if mode == "block":
+                if mode == "block" and parent_node is not None:
                     if id < sibling_id:
                         if node.has_valid_coordinate_values():
-                            if bdr is None:
+                            if bdr is None or first_block_node is None:
                                 bdr = (node.x, node.y, node.x, node.y)
+                                first_block_node = node
+                                bdr_limit = _get_limit_bdr(
+                                    node.x, node.y, 500.0)
                             else:
-                                bdr = (
-                                    min(node.x, bdr[0]),
-                                    min(node.y, bdr[1]),
-                                    max(node.x, bdr[2]),
-                                    max(node.y, bdr[3]),
-                                )
+                                if node.x < bdr_limit[0] or node.y < bdr_limit[1] \
+                                        or node.x > bdr_limit[2] or node.y > bdr_limit[3]:
+                                    # The node is too far from the first node.
+                                    pass
+                                else:
+                                    bdr = (
+                                        min(node.x, bdr[0]),
+                                        min(node.y, bdr[1]),
+                                        max(node.x, bdr[2]),
+                                        max(node.y, bdr[3]),
+                                    )
 
                     if id == sibling_id - 1:
                         mode = ""
                         if bdr:
-                            file_idx.insert(
+                            _insert_node_to_rtree(
                                 id=parent_node.id,
                                 coordinates=bdr,
                             )
-                        else:
-                            # All child nodes have invalid coordinate values
-                            key = (parent_node.x, parent_node.y)
-                            if parent_node.has_valid_coordinate_values() and \
-                                    key not in registered_coordinates:
-                                file_idx.insert(
-                                    id=parent_node.id,
-                                    coordinates=(node.x, node.y,
-                                                 node.x, node.y),
-                                )
-                                registered_coordinates.add(key)
 
                     if id >= sibling_id:
-                        import pdb
-                        pdb.set_trace()
+                        raise RTreeError((
+                            f"データファイルが異常です。"
+                            f"The child node ({id}) of the parent_node ({parent_node.id}) "
+                            f"exceeded the sibling_id:{sibling_id}."))
 
                     continue
 
@@ -425,7 +442,7 @@ class Index(object):
                     if key in registered_coordinates:
                         continue
 
-                    file_idx.insert(
+                    _insert_node_to_rtree(
                         id=id,
                         coordinates=(node.x, node.y, node.x, node.y),
                     )
@@ -441,8 +458,11 @@ class Index(object):
 
                     if node.has_valid_coordinate_values():
                         bdr = (node.x, node.y, node.x, node.y)
+                        first_block_node = node
+                        bdr_limit = _get_limit_bdr(node.x, node.y, 500.0)
                     else:
                         bdr = None
+                        first_block_node = None
 
                     continue
 

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/strlib.py

@@ -1,6 +1,6 @@
 from logging import getLogger
 import re
-from typing import Optional, Union
+from typing import Union
 
 logger = getLogger(__name__)
 

{jageocoder-2.2.0 → jageocoder-2.2.1}/jageocoder/trie.py

@@ -1,6 +1,5 @@
 from logging import getLogger
 import os
-from pathlib import Path
 
 import marisa_trie
 

{jageocoder-2.2.0 → jageocoder-2.2.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "jageocoder"
-version = "2.2.0"
+version = "2.2.1"
 description = "A Japanese-address geocoder for Python."
 authors = [{name = "Takeshi Sagara", email = "sagara@info-proto.com"}]
 readme = "README.md"
@@ -28,7 +28,10 @@ dependencies = [
 
 [project.optional-dependencies]
 dev = [
-    "twine >=5.1.1"
+    "flake8",
+    "twine >=5.1.1",
+    "tox >=4.0.0",
+    "tox-uv"
 ]
 doc = [
     "sphinx-rtd-theme >=1.2.0"