feat: expose num_rows api for python LanceFileReader (#4053)

Co-authored-by: Weston Pace <weston.pace@gmail.com>

Author: yanghua

python/python/lance/file.py

@@ -179,6 +179,10 @@ def read_global_buffer(self, index: int) -> bytes:
         """
         return self._reader.read_global_buffer(index)
 
+    def num_rows(self) -> int:
+        """Return the number of rows belonging to the data file."""
+        return self._reader.num_rows()
+
 
 class LanceFileWriter:
     """

python/python/lance/lance/__init__.pyi

@@ -121,6 +121,7 @@ class LanceFileReader:
     def read_global_buffer(self, index: int) -> bytes: ...
     def metadata(self) -> LanceFileMetadata: ...
     def file_statistics(self) -> LanceFileStatistics: ...
+    def num_rows(self): ...
 
 class LanceBufferDescriptor:
     position: int

python/python/tests/test_file.py

@@ -110,6 +110,17 @@ def test_take(tmp_path):
     assert table == pa.table({"a": [0, 77, 83]})
 
 
+def test_num_rows(tmp_path):
+    path = tmp_path / "foo.lance"
+    schema = pa.schema([pa.field("a", pa.int64())])
+    writer = LanceFileWriter(str(path), schema)
+    writer.write_batch(pa.table({"a": [i for i in range(100)]}))
+    writer.close()
+
+    reader = LanceFileReader(str(path))
+    assert reader.num_rows() == 100
+
+
 def check_round_trip(tmp_path, table):
     path = tmp_path / "foo.lance"
     with LanceFileWriter(str(path), table.schema) as writer:

python/src/file.rs

@@ -553,6 +553,10 @@ impl LanceFileReader {
             .infer_error()?;
         Ok(buffer_bytes.to_vec())
     }
+
+    pub fn num_rows(&mut self) -> u64 {
+        self.inner.num_rows()
+    }
 }
 
 #[cfg(test)]