Merge pull request #241 from Living-with-machines/update_file_saving

Update file saving - save as csv files

Author: rwood-97

docs/source/Worked-examples/mnist_pipeline.ipynb

@@ -1237,7 +1237,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "predictions_df.to_csv(\"./predictions_df.csv\", sep=\"\\t\", index_label=\"image_id\")"
+    "predictions_df.to_csv(\"./predictions_df.csv\", sep=\",\", index_label=\"image_id\")"
    ]
   }
  ],

docs/source/Worked-examples/plant_pipeline.ipynb

@@ -2296,7 +2296,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "predictions_df.to_csv(\"./predictions_df.csv\", sep=\"\\t\", index_label=\"image_id\")"
+    "predictions_df.to_csv(\"./predictions_df.csv\", sep=\",\", index_label=\"image_id\")"
    ]
   },
   {

mapreader/annotate/utils.py

@@ -529,7 +529,7 @@ def prepare_annotation(
                 metadata=annot_file,
                 index_col=0,
                 ignore_mismatch=True,
-                delimiter="\t",
+                delimiter=",",
                 tree_level=tree_level,
             )
 
@@ -584,7 +584,7 @@ def prepare_annotation(
                 metadata=annot_file,
                 index_col=0,
                 ignore_mismatch=True,
-                delimiter="\t",
+                delimiter=",",
                 tree_level=tree_level,
             )
         # convert images to dataframe
@@ -659,7 +659,7 @@ def save_annotation(
 
     # Read an existing annotation file (for the same task and userID)
     try:
-        image_df = pd.read_csv(annot_file, sep="\t", index_col=0)
+        image_df = pd.read_csv(annot_file, index_col=0)
     except:
         image_df = pd.DataFrame(columns=["image_id", "image_path", "label"])
 
@@ -684,7 +684,7 @@ def save_annotation(
 
     if len(image_df) > 0:
         #image_df = image_df.set_index("image_id")
-        image_df.to_csv(annot_file, mode="w", sep="\t")
+        image_df.to_csv(annot_file, mode="w")
         print(f"[INFO] Save {newly_annotated} new annotations to {annot_file}")
         print(f"[INFO] {new_labels} labels were not already stored")
         print(f"[INFO] Total number of saved annotations: {len(image_df)}")

mapreader/classify/datasets.py

@@ -28,7 +28,7 @@ def __init__(
         self,
         patch_df: Union[pd.DataFrame, str],
         transform: Union[str, transforms.Compose, Callable],
-        delimiter: str = "\t",
+        delimiter: str = ",",
         patch_paths_col: Optional[str] = "image_path",
         label_col: Optional[str] = None,
         label_index_col: Optional[str] = None,
@@ -47,7 +47,7 @@ def __init__(
             and performs image transformations can be used.
             At minimum, transform should be ``torchvision.transforms.ToTensor()``.
         delimiter : str, optional
-            The delimiter to use when reading the dataframe. By default ``"\t"``.
+            The delimiter to use when reading the dataframe. By default ``","``.
         patch_paths_col : str, optional
             The name of the column in the DataFrame containing the image paths. Default is "image_path".
         label_col : str, optional
@@ -329,7 +329,7 @@ def __init__(
         patch_df: Union[pd.DataFrame, str],
         transform1: str,
         transform2: str,
-        delimiter: str = "\t",
+        delimiter: str = ",",
         patch_paths_col: Optional[str] = "image_path",
         label_col: Optional[str] = None,
         label_index_col: Optional[str] = None,
@@ -356,7 +356,7 @@ def __init__(
             Torchvision transform to be applied to target images.
             Either "train" or "val".
         delimiter : str
-            The delimiter to use when reading the csv file. By default ``"\t"``.
+            The delimiter to use when reading the csv file. By default ``","``.
         patch_paths_col : str, optional
             The name of the column in the DataFrame containing the image paths. Default is "image_path".
         label_col : str, optional

mapreader/classify/load_annotations.py

@@ -32,7 +32,7 @@ def __init__(self):
     def load(
         self,
         annotations: Union[str, pd.DataFrame],
-        delimiter: Optional[str] = "\t",
+        delimiter: Optional[str] =",",
         id_col: Optional[str] = "image_id",
         patch_paths_col: Optional[str] = "image_path",
         label_col: Optional[str] = "label",
@@ -48,7 +48,7 @@ def load(
             The annotations.
             Can either be the path to a csv file or a pandas.DataFrame.
         delimiter : Optional[str], optional
-            The delimiter to use when loading the csv file as a dataframe, by default "\t".
+            The delimiter to use when loading the csv file as a dataframe, by default ",".
         id_col : Optional[str], optional
             The name of the column which contains the image IDs, by default "image_id".
         patch_paths_col : Optional[str], optional
@@ -123,7 +123,7 @@ def load(
     def _load_annotations_csv(
         self,
         annotations: str,
-        delimiter: Optional[str] = "\t",
+        delimiter: Optional[str] = ",",
         scramble_frame: Optional[bool] = False,
         reset_index: Optional[bool] = False,
     ) -> pd.DataFrame:
@@ -134,7 +134,7 @@ def _load_annotations_csv(
         annotations : str
             The path to the annotations csv file.
         delimiter : Optional[str], optional
-            The delimiter to use when loading the csv file as a dataframe, by default "\t".
+            The delimiter to use when loading the csv file as a dataframe, by default ",".
         scramble_frame : Optional[bool], optional
             Whether to shuffle the rows of the dataframe, by default False.
         reset_index : Optional[bool], optional
@@ -488,8 +488,8 @@ def create_datasets(
             assert len(self.annotations) == len(df_train) + len(df_val) + len(df_test)
 
         else:
-            df_val = labels_temp
-            df_test = None
+            df_val = df_temp
+            df_test = pd.DataFrame(columns=self.annotations.columns)
             assert len(self.annotations) == len(df_train) + len(df_val)
 
         train_dataset = PatchDataset(
@@ -506,13 +506,12 @@ def create_datasets(
             label_col=self.label_col,
             label_index_col="label_index",
         )
-        if df_test is not None:
-            test_dataset = PatchDataset(
-                df_test,
-                test_transform,
-                patch_paths_col=self.patch_paths_col,
-                label_col=self.label_col,
-                label_index_col="label_index",
+        test_dataset = PatchDataset(
+            df_test,
+            test_transform,
+            patch_paths_col=self.patch_paths_col,
+            label_col=self.label_col,
+            label_index_col="label_index",
             )
 
         datasets = {"train": train_dataset, "val": val_dataset, "test": test_dataset}

mapreader/download/sheet_downloader.py

@@ -555,13 +555,13 @@ def _save_metadata(
         ).T
 
         if os.path.exists(out_filepath):
-            existing_metadata_df = pd.read_csv(out_filepath, sep="\t", index_col=0)
+            existing_metadata_df = pd.read_csv(out_filepath, sep=",", index_col=0)
             metadata_df = pd.concat([existing_metadata_df, new_metadata_df], ignore_index=True)
             metadata_df.drop_duplicates(subset=["grid_bb"], keep="first", inplace=True)
         else: 
             metadata_df = new_metadata_df
 
-        metadata_df.to_csv(out_filepath, sep="\t")
+        metadata_df.to_csv(out_filepath, sep=",")
 
     def _download_map_sheets(self, features: list, path_save: Optional[str] = "maps", metadata_fname: Optional[str] = "metadata.csv", overwrite: Optional[bool] = False):
         """Download map sheets from a list of features.

mapreader/load/images.py

@@ -260,7 +260,7 @@ def add_metadata(
         self,
         metadata: Union[str, pd.DataFrame],
         index_col: Optional[Union[int, str]] = 0,
-        delimiter: Optional[str] = "\t",
+        delimiter: Optional[str] = ",",
         columns: Optional[List[str]] = None,
         tree_level: Optional[str] = "parent",
         ignore_mismatch: Optional[bool] = False,
@@ -271,7 +271,7 @@ def add_metadata(
         Parameters
         ----------
         metadata : str or pandas.DataFrame
-            Path to a ``csv``, ``xls`` or ``xlsx`` file or a pandas DataFrame that contains the metadata information.
+            Path to a ``csv`` (or similar), ``xls`` or ``xlsx`` file or a pandas DataFrame that contains the metadata information.
         index_col : int or str, optional
             Column to use as the index when reading the file and converting into a panda.DataFrame.
             Accepts column indices or column names.
@@ -280,7 +280,7 @@ def add_metadata(
             Only used if a file path is provided as the ``metadata`` parameter.
             Ignored if ``columns`` parameter is passed.
         delimiter : str, optional
-            Delimiter used in the ``csv`` file, by default ``"\t"``.
+            Delimiter used in the ``csv`` file, by default ``","``.
 
             Only used if a ``csv`` file path is provided as
             the ``metadata`` parameter.
@@ -323,31 +323,32 @@ def add_metadata(
 
         else: #if not df
             if os.path.isfile(metadata):
-                if metadata.endswith('csv'):
+                if metadata.endswith(('xls', 'xlsx')):
                     if columns:
-                        metadata_df = pd.read_csv(
-                            metadata, usecols=columns, delimiter=delimiter
+                        metadata_df = pd.read_excel(
+                            metadata, usecols=columns,
                             )
                     else:
-                        metadata_df = pd.read_csv(
-                            metadata, index_col=index_col, delimiter=delimiter
+                        metadata_df = pd.read_excel(
+                            metadata, index_col=index_col,
                             )
                         columns=list(metadata_df.columns)
-                
-                elif metadata.endswith(('xls', 'xlsx')):
+            
+                elif metadata.endswith('sv'): #csv, tsv, etc
                     if columns:
-                        metadata_df = pd.read_excel(
-                            metadata, usecols=columns,
+                        metadata_df = pd.read_csv(
+                            metadata, usecols=columns, delimiter=delimiter
                             )
                     else:
-                        metadata_df = pd.read_excel(
-                            metadata, index_col=index_col,
+                        metadata_df = pd.read_csv(
+                            metadata, index_col=index_col, delimiter=delimiter
                             )
                         columns=list(metadata_df.columns)
+                
 
             else:
                 raise ValueError(
-                    "[ERROR] ``metadata`` should either be the path to a ``csv``, ``xls`` or ``xlsx`` file or a pandas DataFrame."  # noqa
+                    "[ERROR] ``metadata`` should either be the path to a ``csv`` (or similar), ``xls`` or ``xlsx`` file or a pandas DataFrame."  # noqa
                 )
 
         # identify image_id column
@@ -1225,7 +1226,12 @@ def calc_pixel_stats(
                         # Calculate std pixel values
                         self.patches[patch][f"std_pixel_{band}"] = img_std[i] / 255
 
-    def convert_images(self, save: Optional[bool] = False, save_format: Optional[str] ="csv") -> Tuple[pd.DataFrame, pd.DataFrame]:
+    def convert_images(
+            self, 
+            save: Optional[bool] = False, 
+            save_format: Optional[str] ="csv",
+            delimiter: Optional[str]=",",
+    ) -> Tuple[pd.DataFrame, pd.DataFrame]:
         """
         Convert the ``MapImages`` instance's ``images`` dictionary into pandas
         DataFrames for easy manipulation.
@@ -1239,6 +1245,8 @@ def convert_images(self, save: Optional[bool] = False, save_format: Optional[str
             If ``save = True``, the file format to use when saving the dataframes.
             Options of csv ("csv") or excel ("excel" or "xlsx"). 
             By default, "csv".
+        delimiter : str, optional
+            The delimiter to use when saving the dataframe. By default ``","``.
 
         Returns
         -------
@@ -1255,9 +1263,9 @@ def convert_images(self, save: Optional[bool] = False, save_format: Optional[str
         if save:
 
             if save_format == "csv":
-                parent_df.to_csv("parent_df.csv", sep="\t")
+                parent_df.to_csv("parent_df.csv", sep=delimiter)
                 print('[INFO] Saved parent dataframe as "parent_df.csv"')
-                patch_df.to_csv("patch_df.csv", sep="\t")
+                patch_df.to_csv("patch_df.csv", sep=delimiter)
                 print('[INFO] Saved patch dataframe as "patch_df.csv"')
             elif save_format in ["excel", "xlsx"]:
                 parent_df.to_excel("parent_df.xlsx")
@@ -1872,6 +1880,7 @@ def load_csv(
         clear_images: Optional[bool] = False,
         index_col_patch: Optional[int] = 0,
         index_col_parent: Optional[int] = 0,
+        delimiter: Optional[str] = ",",
     ) -> None:
         """
         Load CSV files containing information about parent and patches,
@@ -1891,6 +1900,8 @@ def load_csv(
             Column to set as index for the patch DataFrame, by default ``0``.
         index_col_parent : int, optional
             Column to set as index for the parent DataFrame, by default ``0``.
+        delimiter : str, optional
+            The delimiter to use when reading the dataframe. By default ``","``.
 
         Returns
         -------
@@ -1905,12 +1916,12 @@ def load_csv(
             raise ValueError("[ERROR] Please pass ``patch_path`` as string.")
 
         if os.path.isfile(parent_path):
-            parent_df = pd.read_csv(parent_path, index_col=index_col_parent)
+            parent_df = pd.read_csv(parent_path, index_col=index_col_parent, sep=delimiter)
         else:
             raise ValueError(f"[ERROR] {parent_path} cannot be found.")
 
         if os.path.isfile(patch_path):
-            patch_df = pd.read_csv(patch_path, index_col=index_col_patch)
+            patch_df = pd.read_csv(patch_path, index_col=index_col_patch, sep=delimiter)
         else:
             raise ValueError(f"[ERROR] {patch_path} cannot be found.")