Methods for mapping Wikipedia ID (Page ID) to Wikipedia page title an…

…d back (tests included) (#15)

* added methods for mapping wikipedia id (page id) to wikidata id and back

README.rst

@@ -131,6 +131,28 @@ Map Wikidata id to Wikipedia id
 Mapping Wikidata id to Wikipedia id can lead to more than one result, as some pages in Wikipedia are
 redirects, all linking to the same Wikidata item.
 
+Map Wikipedia id to Wikipedia page title
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. code:: python
+
+    from wikimapper import WikiMapper
+
+    mapper = WikiMapper("index_enwiki-latest.db")
+    page_title = mapper.wikipedia_id_to_title(3342)
+    print(page_title)  # Bundesrepublik_Deutschland
+
+Map Wikipedia page title to Wikipedia id
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+.. code:: python
+
+    from wikimapper import WikiMapper
+
+    mapper = WikiMapper("index_enwiki-latest.db")
+    wikipedia_id = mapper.title_to_wikipedia_id("Germany")
+    print(wikipedia_id)  # 11867
+
 Create your own index
 ~~~~~~~~~~~~~~~~~~~~~
 

tests/test_mapper.py

@@ -94,3 +94,58 @@ def test_id_to_wikipedia_ids(bavarian_wiki_mapper, wikidata_id: str, expected: L
     wikipedia_ids = mapper.id_to_wikipedia_ids(wikidata_id)
 
     assert set(wikipedia_ids) == set(expected)
+
+
+@pytest.mark.parametrize(
+    "wikipedia_id, expected",
+    [
+        (24520, "Stoaboog"),
+        (8535, "Wechslkrod"),
+        (32218, None),  # Wechslkrod, but namespace 1, so cannot be in the database
+        (32176, "Wickiana"),
+        (32252, "Ulrich_Zwingli"),
+        (32311, "Jingstes_Gricht"),
+        (2143, "Sånkt_Johann_im_Pongau"),
+        (2217, "Quadrátkilometa"),
+        (4209, "D'_boarische_Woocha"),
+        (1997, "Brezn"),
+        (5740, None),  # Brezn, but namespace 1, so cannot be in the database
+        (24100, "Brezel"),
+        (28193, "Brezen"),
+        (105208, "Vulkanologie"),
+        (105288, "Vuikanologie"),
+    ]
+)
+def test_wikipedia_id_to_title(bavarian_wiki_mapper, wikipedia_id: int, expected: str):
+    mapper = bavarian_wiki_mapper
+
+    title = mapper.wikipedia_id_to_title(wikipedia_id)
+
+    assert title == expected
+
+
+@pytest.mark.parametrize(
+    "title, expected",
+    [
+        ("Stoaboog", 24520),
+        ("Wechslkrod", 8535),
+        ("Wickiana", 32176),
+        ("Ulrich_Zwingli", 32252),
+        ("Jingstes_Gricht", 32311),
+        ("Sånkt_Johann_im_Pongau", 2143),
+        ("Quadrátkilometa", 2217),
+        ("D'_boarische_Woocha", 4209),
+        ("Brezn", 1997),
+        ("Brezel", 24100),
+        ("Brezen", 28193),
+        ("Vulkanologie", 105208),
+        ("Vuikanologie", 105288),
+        ("xxxxxxxxxx", None),
+    ]
+)
+def test_wikipedia_id_to_title(bavarian_wiki_mapper, title: str, expected: int):
+    mapper = bavarian_wiki_mapper
+
+    wikipedia_id = mapper.title_to_wikipedia_id(title)
+
+    assert wikipedia_id == expected

wikimapper/mapper.py

@@ -114,3 +114,46 @@ def id_to_wikipedia_ids(self, wikidata_id: str) -> List[int]:
         results = c.fetchall()
 
         return [e[0] for e in results]
+
+    def wikipedia_id_to_title(self, wikipedia_id: int) -> Optional[str]:
+        """Given a Wikipedia ID (in other words Page ID), returns the corresponding page title.
+
+        Args:
+            wikipedia_id (int): The Wikipedia ID to map, e.g. `11867`
+
+        Returns:
+            Optional[str]: If a mapping found for `wikipedia_id`, then return
+                           it, else return `None`.
+        """
+
+        c = self.conn.execute(
+            "SELECT wikipedia_title FROM mapping WHERE wikipedia_id=?", (wikipedia_id,)
+        )
+        result = c.fetchone()
+
+        if result is not None and result[0] is not None:
+            return result[0]
+        else:
+            return None
+
+    def title_to_wikipedia_id(self, page_title: str) -> Optional[int]:
+        """Given a Wikipedia page title, returns the corresponding Wikipedia id.
+
+        Args:
+            page_title (str): The Wikipedia page title to map, e.g. `Germany`
+
+        Returns:
+            Optional[str]: If a mapping found for `page_title`, then return
+                           it, else return `None`.
+        """
+
+        # no need for `DISTINCT` as `wikipedia_id` is a PRIMARY KEY, thus we have no duplicates there
+        c = self.conn.execute(
+            "SELECT wikipedia_id FROM mapping WHERE wikipedia_title=?", (page_title,)
+        )
+        result = c.fetchone()
+
+        if result is not None and result[0] is not None:
+            return result[0]
+        else:
+            return None