Pandas support

pyproject.toml

@@ -18,6 +18,7 @@ repository = "https://github.com/biomarkersParkinson/tsdf"
 [tool.poetry.dependencies]
 python = "^3.9"
 numpy = "^1.24.1"
+pandas = "^2.1.3"
 
 [tool.poetry.group.dev.dependencies]
 pytest = "^7.2.0"

src/tsdf/__init__.py

@@ -16,9 +16,11 @@
 
 from .write_binary import (
     write_binary_file,
+    write_dataframe_to_binaries,
 )
 from .read_binary import (
     load_binary_from_metadata,
+    load_binaries_to_dataframe,
 )
 
 from .tsdfmetadata import TSDFMetadata
@@ -31,7 +33,9 @@
     'load_metadata_legacy_file',
     'write_metadata',
     'write_binary_file',
+    'write_dataframe_to_binaries',
     'load_binary_from_metadata',
+    'load_binaries_to_dataframe',
     'TSDFMetadata',
     'constants'
 ]

src/tsdf/read_binary.py

@@ -6,10 +6,30 @@
 
 import os
 import numpy as np
+import pandas as pd
 from tsdf import numpy_utils 
 from tsdf import tsdfmetadata
 
 
+def load_binaries_to_dataframe(metadatas: '[tsdfmetadata.TSDFMetadata]') -> pd.DataFrame:
+    """
+    Load binary files associated with TSDF and return a combined pandas DataFrame.
+
+    :param metadatas: list of TSDFMetadata objects.
+
+    :return: pandas DataFrame containing the combined data.
+    """
+    # Load the data
+    dataFrames = []
+    for metadata in metadatas:
+        data = load_binary_from_metadata(metadata)
+        df = pd.DataFrame(data, columns=metadata.channels)
+        dataFrames.append(df)
+
+    # Merge the data
+    return pd.concat(dataFrames, axis=1)
+
+
 def load_binary_from_metadata(
     metadata: 'tsdfmetadata.TSDFMetadata', start_row: int = 0, end_row: int = -1
 ) -> np.ndarray:

src/tsdf/tsdfmetadata.py

@@ -50,7 +50,7 @@ class TSDFMetadata:
 
     def __init__(
         self, dictionary: Dict[str, Any], dir_path: str, file_name: str = ""
-    ) -> None:
+    ) -> None: #TODO: why not get file_name from dictionary? Which one is used; is this clear?
         """
         The default constructor takes a dictionary as an argument and creates each
         field as a separate property.\\
@@ -60,7 +60,7 @@ def __init__(
         :param dir_path: path to the directory where the metadata file is stored.
         :param file_name: (optional) name of the metadata file.
         """
-        parse_metadata.contains_tsdf_mandatory_fields(dictionary)
+        parse_metadata.contains_tsdf_mandatory_fields(dictionary) #TODO: how to load a dict that is not complete yet?
         for key, value in dictionary.items():
             setattr(self, key, value)
         self.file_dir_path = dir_path

src/tsdf/write_binary.py

@@ -7,11 +7,38 @@
 import os
 from typing import Any, Dict
 import numpy as np
+import pandas as pd
 from tsdf import numpy_utils 
 
 from tsdf.tsdfmetadata import TSDFMetadata
 
 
+def write_dataframe_to_binaries(
+    file_dir: str, df: pd.DataFrame, metadatas: [TSDFMetadata]
+) -> None:
+    """
+    Save binary file based on the provided pandas DataFrame.
+
+    :param file_dir: path to the directory where the file will be saved.
+    :param file_name: name of the file to be saved.
+    :param data: pandas DataFrame containing the data.
+    :param metadata: dictionary containing the metadata.
+
+    :return: TSDFMetadata object.
+    """
+    for metadata in metadatas:
+        file_name = metadata.file_name
+        path = os.path.join(file_dir, file_name)
+
+        # Write
+        data = df[metadata.channels].to_numpy() # TODO: derive channels from dataframe or use specified in metadata? Also for file_name?
+        data.tofile(path)
+
+        # Update metadata with data properties
+        data_props = _get_metadata_from_ndarray(data)
+        for key in data_props:
+            metadata.__setattr__(key, data_props[key])
+
 
 def _get_metadata_from_ndarray(data: np.ndarray) -> Dict[str, Any]:
     """
@@ -48,5 +75,6 @@ def write_binary_file(
     data.tofile(path)
     metadata.update(_get_metadata_from_ndarray(data))
     metadata.update({"file_name": file_name})
+    #TODO: update file_dir_path?
 
     return TSDFMetadata(metadata, file_dir)

tests/test_read_binary.py

@@ -49,3 +49,9 @@ def test_random_access(shared_datadir):
     data = tsdf.load_binary_from_metadata(metadata[name + ".bin"], 2, 6)
     assert(data.shape == (4, 3))
     assert(data.dtype == "int16")
+
+def test_load_binary_to_dataframe(shared_datadir):
+    metadata = tsdf.load_metadata_from_path(shared_datadir / "ppp_format_meta.json")
+    df = tsdf.load_binaries_to_dataframe([metadata["ppp_format_time.bin"], metadata["ppp_format_samples.bin"]])
+    assert(df.shape == (17, 7))
+    assert(df.columns.tolist() == ["time", "acceleration_x", "acceleration_y", "acceleration_z", "rotation_x", "rotation_y", "rotation_z"])