Add category_id_start parameter to support starting category IDs from…

… 1 (#29)

* Included:
If data is structured in subdirectories, the pathes are missing in the "file_name" attribute. Instead of subfolder_1/image.jpeg the filename contains only image.jpeg.

New:
Add category_id_start parameter to support starting category IDs from 1

* Added missing stuff

* Incremented version number. Reverted to fire.Fire(convert)

README.md

@@ -39,6 +39,10 @@ labelme2coco path/to/labelme/dir
 labelme2coco path/to/labelme/dir --train_split_rate 0.85
 ```
 
+```python
+labelme2coco path/to/labelme/dir --category_id_start 1
+```
+
 ### Advanced Usage
 
 ```python
@@ -54,8 +58,11 @@ export_dir = "tests/data/"
 # set train split rate
 train_split_rate = 0.85
 
+# set category ID start value
+category_id_start = 1
+
 # convert labelme annotations to coco
-labelme2coco.convert(labelme_folder, export_dir, train_split_rate)
+labelme2coco.convert(labelme_folder, export_dir, train_split_rate, category_id_start=category_id_start)
 ```
 
 ```python
@@ -71,14 +78,17 @@ labelme_val_folder = "tests/data/labelme_annot"
 # set path for coco json to be saved
 export_dir = "tests/data/"
 
+# set category ID start value
+category_id_start = 1
+
 # create train coco object
-train_coco = get_coco_from_labelme_folder(labelme_train_folder)
+train_coco = get_coco_from_labelme_folder(labelme_train_folder, category_id_start=category_id_start)
 
 # export train coco json
 save_json(train_coco.json, export_dir+"train.json")
 
 # create val coco object
-val_coco = get_coco_from_labelme_folder(labelme_val_folder, coco_category_list=train_coco.json_categories)
+val_coco = get_coco_from_labelme_folder(labelme_val_folder, coco_category_list=train_coco.json_categories, category_id_start=category_id_start)
 
 # export val coco json
 save_json(val_coco.json, export_dir+"val.json")

labelme2coco/__init__.py

@@ -1,6 +1,6 @@
 from __future__ import absolute_import
 
-__version__ = "0.2.5"
+__version__ = "0.2.6"
 
 import logging
 import os
@@ -25,14 +25,16 @@ def convert(
     export_dir: str = "runs/labelme2coco/",
     train_split_rate: float = 1,
     skip_labels: List[str] = [],
+    category_id_start: int = 0,
 ):
     """
     Args:
         labelme_folder: folder that contains labelme annotations and image files
         export_dir: path for coco jsons to be exported
         train_split_rate: ration fo train split
+        category_id_start: starting value for category IDs (default: 0)
     """
-    coco = get_coco_from_labelme_folder(labelme_folder, skip_labels=skip_labels)
+    coco = get_coco_from_labelme_folder(labelme_folder, skip_labels=skip_labels, category_id_start=category_id_start)
     if train_split_rate < 1:
         result = coco.split_coco_as_train_val(train_split_rate)
         # export train split

labelme2coco/labelme2coco.py

@@ -15,7 +15,7 @@ def __init__(self):
 
 
 def get_coco_from_labelme_folder(
-    labelme_folder: str, coco_category_list: List = None, skip_labels: List[str] = []
+    labelme_folder: str, coco_category_list: List = None, skip_labels: List[str] = [], category_id_start: int = 0
 ) -> Coco:
     """
     Args:
@@ -37,19 +37,21 @@ def get_coco_from_labelme_folder(
         print(f"Will skip the following annotated labels: {skip_labels}")
 
     # parse labelme annotations
-    category_ind = 0
+    # depending on cli arguments, will start counting at 1
+    category_ind = category_id_start
     for json_path in tqdm(
         labelme_json_list, "Converting labelme annotations to COCO format"
     ):
+        # Taken from https://github.com/fcakyon/labelme2coco/pull/17
         data = load_json(json_path)
         # get image size
-        image_path = str(Path(labelme_folder) / data["imagePath"])
+        image_path = str(Path(json_path).parent / data["imagePath"])
         # use the image sizes provided by labelme (they already account for
         # things such as EXIF orientation)
         width = data["imageWidth"]
         height = data["imageHeight"]
         # init coco image
-        coco_image = CocoImage(file_name=data["imagePath"], height=height, width=width)
+        coco_image = CocoImage(file_name=image_path, height=height, width=width)
         # iterate over annotations
         for shape in data["shapes"]:
             # set category name and id
@@ -120,4 +122,4 @@ def get_coco_from_labelme_folder(
 
 if __name__ == "__main__":
     labelme_folder = "tests/data/labelme_annot"
-    coco = get_coco_from_labelme_folder(labelme_folder)
+    coco = get_coco_from_labelme_folder(labelme_folder)