Add import EPRI dataset script

Author: Vitalii Bulyzhyn

sdk/samples/EPRI dataset import/.gitignore

@@ -0,0 +1,4 @@
+Overhead-Distribution-Labels.csv
+images/*.JPG
+.env
+.DS_store

sdk/samples/EPRI dataset import/Readme.md

@@ -0,0 +1,43 @@
+## Description
+
+This is a simple script to import [EPRI Distribution Inspection Imagery](https://www.kaggle.com/datasets/dexterlewis/epri-distribution-inspection-imagery) to Diffgram
+
+So far we, it only imports annotation with type **polygon** and skips annotations with type **line** (polyline type of instance doesn't exist on Diffgram yet), but that will be improved in future version
+
+## Usage
+
+1. Create and activate virtual environment
+
+```
+virtualenv your-env-name
+source your-env-name/bin/activate
+```
+
+2. Install dependencies from **requirements.txt**:
+
+```
+pip install -r requirements.txt
+```
+
+3. Download [annotations file](https://publicstorageaccnt.blob.core.windows.net/drone-distribution-inspection-imagery/Overhead-Distribution-Labels.csv) and place it to the root folder
+
+4. Download images and unzip them to **images** folder
+
+5. Create .env file and set environmental variables:
+
+```
+touch .env
+```
+
+```
+PROJECT_STRING_ID=project-string-id
+CLIENT_ID=client-id
+CLIENT_SECRET=client-secret
+HOST=https://example.com
+```
+
+6. Run script:
+
+```
+python import.py
+```

sdk/samples/EPRI dataset import/images/.gitkeep

@@ -0,0 +1,108 @@
+import os
+import ast
+import pandas as pd
+from dotenv import load_dotenv
+from diffgram import Project
+
+load_dotenv()
+
+image_list = os.listdir('images')
+
+project = Project(
+    project_string_id = os.getenv('PROJECT_STRING_ID'),
+    client_id = os.getenv('CLIENT_ID'),
+    client_secret = os.getenv('CLIENT_SECRET'),
+    host = os.getenv('HOST')
+)
+
+number_of_images = None
+while number_of_images is None:
+    try:
+        number_of_images_to_import = input("How many images do you want to import? (blank to import all) ")
+        if not number_of_images_to_import:
+            number_of_images = len(image_list)
+        number_of_images = int(number_of_images_to_import)
+    except:
+        print("Invalid input: please input positive number")
+
+image_list = image_list[:int(number_of_images_to_import)]
+
+new_schema_name = None
+while True:
+    try:
+        new_schema_name = input("Give a name for a new schema: ")
+        schema = project.new_schema(name=new_schema_name)
+        print("Schema successfully created")
+        break
+    except:
+        print("Seems like schema with this name already exists")
+
+df = None
+while True:
+    try:
+        annotation_file_name = input("What is the name of the file with annotations? (leave blank to use default Overhead-Distribution-Labels.csv)")
+        if not annotation_file_name:
+            df = pd.read_csv ('Overhead-Distribution-Labels.csv')
+            break
+        df = pd.read_csv (annotation_file_name)
+        break
+    except:
+        print("Seems like annotation file is not here")
+
+imported_label_traker = []
+lables_objects = {}
+
+succeslully_imported = []
+import_errors = []
+
+for image in image_list:
+    image_relate_df = df[df['External ID'] == image]
+    labels = image_relate_df['Label']
+    external_id = image_relate_df['External ID']
+
+    instance_list = []
+
+    for label in labels:
+        label_dict = ast.literal_eval(label)
+
+        for object in label_dict['objects']:
+            label = {}
+
+            if object['value'] not in imported_label_traker:
+                label = project.label_new({'name': object['value']}, schema.get('id'))
+                lables_objects[label['label']['name']] = label
+            else:
+                label = lables_objects[object['value']]
+
+            polygone = object.get('polygon')
+            line = object.get('line')
+
+            if polygone:
+                instance_list.append({
+                    "type": 'polygon',
+                    "points": polygone,
+                    "label_file_id": label['id']
+                })
+            elif line:
+                pass
+            else:
+                pass
+
+            imported_label_traker.append(object['value'])
+        
+        try:
+            result = project.file.from_local(
+                path=f'./images/{image}', 
+                instance_list = instance_list,
+                convert_names_to_label_files=False
+            )
+
+            succeslully_imported.append(image)
+
+            print(f'{image} has been imported with {len(instance_list)} annotation(s)')
+        except:
+            import_errors.append(image)
+            print(f'Error ocurred while importing {image}')
+
+print(f"Successfully imported {len(succeslully_imported)} file(s): ", succeslully_imported)
+print(f"Errors while importing {len(succeslully_imported)} file(s): ", import_errors)

sdk/samples/EPRI dataset import/import.py

@@ -0,0 +1,15 @@
+certifi==2022.6.15
+charset-normalizer==2.1.0
+diffgram==0.8.5
+idna==3.3
+imageio==2.19.5
+numpy==1.23.1
+pandas==1.4.3
+Pillow==9.2.0
+python-dateutil==2.8.2
+python-dotenv==0.20.0
+pytz==2022.1
+requests==2.28.1
+scipy==1.8.1
+six==1.16.0
+urllib3==1.26.10