Merge pull request #109 from roboflow/login

Quickstart Backbones

Author: Jacobsolawetz

requirements.txt

@@ -17,4 +17,4 @@ wget
 tqdm>=4.41.0
 PyYAML>=5.3.1
 wget
-requests_toolbelt
+requests_toolbelt

roboflow/__init__.py

@@ -1,14 +1,17 @@
 import json
+import os
 import sys
 import time
+from urllib.parse import urlparse
 
 import requests
 
-from roboflow.config import API_URL, APP_URL, DEMO_KEYS
+from roboflow.config import API_URL, APP_URL, DEMO_KEYS, load_roboflow_api_key
 from roboflow.core.project import Project
 from roboflow.core.workspace import Workspace
+from roboflow.util.general import write_line
 
-__version__ = "0.2.34"
+__version__ = "1.0.0"
 
 
 def check_key(api_key, model, notebook, num_retries=0):
@@ -67,14 +70,118 @@ def auth(api_key):
     return Roboflow(api_key, w)
 
 
+def login(workspace=None, force=False):
+    conf_location = os.getenv(
+        "ROBOFLOW_CONFIG_DIR",
+        default=os.getenv("HOME") + "/.config/roboflow/config.json",
+    )
+
+    if os.path.isfile(conf_location) and not force:
+        write_line(
+            "You are already logged into Roboflow. To make a different login, run roboflow.login(force=True)."
+        )
+        return None
+        # we could eventually return the workspace object here
+        # return Roboflow().workspace()
+    elif os.path.isfile(conf_location) and force:
+        os.remove(conf_location)
+
+    if workspace is None:
+        write_line(
+            "visit " + APP_URL + "/auth-cli" " to get your authentication token."
+        )
+    else:
+        write_line(
+            "visit "
+            + APP_URL
+            + "/auth-cli/?workspace="
+            + workspace
+            + " to get your authentication token."
+        )
+
+    token = input("Paste the authentication here token here: ")
+
+    r_login = requests.get(APP_URL + "/query/cliAuthToken/" + token)
+
+    if r_login.status_code == 200:
+        r_login = r_login.json()
+
+        # make config directory if it doesn't exist
+        if not os.path.exists(os.path.dirname(conf_location)):
+            os.mkdir(os.path.dirname(conf_location))
+
+        r_login = {"workspaces": r_login}
+        # set first workspace as default workspace
+
+        default_workspace_id = list(r_login["workspaces"].keys())[0]
+        workspace = r_login["workspaces"][default_workspace_id]
+        r_login["RF_WORKSPACE"] = workspace["url"]
+
+        # write config file
+        with open(conf_location, "w") as f:
+            json.dump(r_login, f, indent=2)
+
+    else:
+        r_login.raise_for_status()
+
+    return None
+    # we could eventually return the workspace object here
+    # return Roboflow().workspace()
+
+
+active_workspace = None
+
+
+def initialize_roboflow():
+    global active_workspace
+
+    conf_location = os.getenv(
+        "ROBOFLOW_CONFIG_DIR",
+        default=os.getenv("HOME") + "/.config/roboflow/config.json",
+    )
+
+    if not os.path.isfile(conf_location):
+        raise RuntimeError(
+            "To use this method, you must first login - run roboflow.login()"
+        )
+    else:
+        if active_workspace == None:
+            active_workspace = Roboflow().workspace()
+
+        return active_workspace
+
+
+def load_model(model_url):
+    operate_workspace = initialize_roboflow()
+
+    if "universe.roboflow.com" in model_url or "app.roboflow.com" in model_url:
+        parsed_url = urlparse(model_url)
+        path_parts = parsed_url.path.split("/")
+        project = path_parts[2]
+        version = int(path_parts[-1])
+    else:
+        raise (
+            "Model URL must be from either app.roboflow.com or universe.roboflow.com"
+        )
+
+    project = operate_workspace.project(project)
+    version = project.version(version)
+    model = version.model
+    return model
+
+
+# continue distributing this object for back compatibility
 class Roboflow:
     def __init__(
         self,
-        api_key="YOUR ROBOFLOW API KEY HERE",
+        api_key=None,
         model_format="undefined",
         notebook="undefined",
     ):
         self.api_key = api_key
+        if self.api_key == None:
+            self.api_key = load_roboflow_api_key()
+
         self.model_format = model_format
         self.notebook = notebook
         self.onboarding = False

roboflow/config.py

@@ -1,26 +1,66 @@
+import json
 import os
 
+
+def get_conditional_configuration_variable(key, default):
+    """Retrieves the configuration variable conditionally.
+        ##1. check if variable is in environment
+        ##2. check if variable is in config file
+        ##3. return default value
+    Args:
+        key (string): The name of the configuration variable.
+        default (string): The default value of the configuration variable.
+    Returns:
+        string: The value of the conditional configuration variable.
+    """
+
+    # default configuration location
+    conf_location = os.getenv(
+        "ROBOFLOW_CONFIG_DIR",
+        default=os.getenv("HOME") + "/.config/roboflow/config.json",
+    )
+
+    # read config file for roboflow if logged in from python or CLI
+    if os.path.exists(conf_location):
+        with open(conf_location) as f:
+            config = json.load(f)
+    else:
+        config = {}
+
+    if os.getenv(key) != None:
+        return os.getenv(key)
+    elif key in config.keys():
+        return config[key]
+    else:
+        return default
+
+
 CLASSIFICATION_MODEL = os.getenv("CLASSIFICATION_MODEL", "ClassificationModel")
 INSTANCE_SEGMENTATION_MODEL = "InstanceSegmentationModel"
 OBJECT_DETECTION_MODEL = os.getenv("OBJECT_DETECTION_MODEL", "ObjectDetectionModel")
 SEMANTIC_SEGMENTATION_MODEL = "SemanticSegmentationModel"
 PREDICTION_OBJECT = os.getenv("PREDICTION_OBJECT", "Prediction")
 
-API_URL = os.getenv("API_URL", "https://api.roboflow.com")
-APP_URL = os.getenv("APP_URL", "https://app.roboflow.com")
-UNIVERSE_URL = os.getenv("UNIVERSE_URL", "https://universe.roboflow.com")
-INSTANCE_SEGMENTATION_URL = os.getenv(
+API_URL = get_conditional_configuration_variable("API_URL", "https://api.roboflow.com")
+APP_URL = get_conditional_configuration_variable("APP_URL", "https://app.roboflow.com")
+UNIVERSE_URL = get_conditional_configuration_variable(
+    "UNIVERSE_URL", "https://universe.roboflow.com"
+)
+
+INSTANCE_SEGMENTATION_URL = get_conditional_configuration_variable(
     "INSTANCE_SEGMENTATION_URL", "https://outline.roboflow.com"
 )
-SEMANTIC_SEGMENTATION_URL = os.getenv(
+SEMANTIC_SEGMENTATION_URL = get_conditional_configuration_variable(
     "SEMANTIC_SEGMENTATION_URL", "https://segment.roboflow.com"
 )
-OBJECT_DETECTION_URL = os.getenv(
+OBJECT_DETECTION_URL = get_conditional_configuration_variable(
     "SEMANTIC_SEGMENTATION_URL", "https://detect.roboflow.com"
 )
 
-CLIP_FEATURIZE_URL = os.getenv("CLIP_FEATURIZE_URL", "CLIP FEATURIZE URL NOT IN ENV")
-OCR_URL = os.getenv("OCR_URL", "OCR URL NOT IN ENV")
+CLIP_FEATURIZE_URL = get_conditional_configuration_variable(
+    "CLIP_FEATURIZE_URL", "CLIP FEATURIZE URL NOT IN ENV"
+)
+OCR_URL = get_conditional_configuration_variable("OCR_URL", "OCR URL NOT IN ENV")
 
 DEMO_KEYS = ["coco-128-sample", "chess-sample-only-api-key"]
 
@@ -30,3 +70,25 @@
 TYPE_SEMANTIC_SEGMENTATION = "semantic-segmentation"
 
 DEFAULT_BATCH_NAME = "Pip Package Upload"
+
+RF_WORKSPACES = get_conditional_configuration_variable("workspaces", default={})
+
+
+def load_roboflow_api_key():
+    RF_WORKSPACE = get_conditional_configuration_variable("RF_WORKSPACE", default=None)
+    RF_WORKSPACES = get_conditional_configuration_variable("workspaces", default={})
+
+    # DEFAULT_WORKSPACE = get_conditional_configuration_variable("default_workspace", default=None)
+    if RF_WORKSPACE == None:
+        RF_API_KEY = None
+    else:
+        RF_API_KEY = None
+        for k in RF_WORKSPACES.keys():
+            workspace = RF_WORKSPACES[k]
+            if workspace["url"] == RF_WORKSPACE:
+                RF_API_KEY = workspace["apiKey"]
+    # ENV API_KEY OVERRIDE
+    if os.getenv("RF_API_KEY") != None:
+        RF_API_KEY = os.getenv("RF_API_KEY")
+
+    return RF_API_KEY

roboflow/core/project.py

@@ -94,6 +94,7 @@ def versions(self):
                 workspace=self.__workspace,
                 project=self.__project_name,
                 public=self.public,
+                colors=self.colors,
             )
             version_array.append(version_object)
         return version_array
@@ -192,12 +193,14 @@ def train(
         },
         speed=None,
         checkpoint=None,
+        plot_in_notebook=False,
     ) -> bool:
         """
         Ask the Roboflow API to train a previously exported version's dataset.
         Args:
             speed: Whether to train quickly or accurately. Note: accurate training is a paid feature. Default speed is `fast`.
             checkpoint: A string representing the checkpoint to use while training
+            plot: Whether to plot the training loss curve. Default is False.
         Returns:
             True
             RuntimeError: If the Roboflow API returns an error with a helpful JSON body
@@ -206,9 +209,11 @@ def train(
 
         new_version = self.generate_version(settings=new_version_settings)
         new_version = self.version(new_version)
-        new_version.train(speed=speed, checkpoint=checkpoint)
+        new_model = new_version.train(
+            speed=speed, checkpoint=checkpoint, plot_in_notebook=plot_in_notebook
+        )
 
-        return True
+        return new_model
 
     def version(self, version_number, local=None):
         """Retrieves information about a specific version, and throws it into an object.
@@ -251,6 +256,7 @@ def version(self, version_number, local=None):
                     workspace=self.__workspace,
                     project=self.__project_name,
                     public=self.public,
+                    colors=self.colors,
                 )
                 return vers
 
@@ -285,7 +291,7 @@ def __image_upload(
             # Construct URL for local image upload
             self.image_upload_url = "".join(
                 [
-                    "https://api.roboflow.com/dataset/",
+                    API_URL + "/dataset/",
                     project_name,
                     "/upload",
                     "?api_key=",
@@ -371,6 +377,9 @@ def __annotation_upload(self, annotation_path: str, image_id: str):
             }
 
         # Set annotation upload url
+
+        project_name = self.id.rsplit("/")[1]
+
         self.annotation_upload_url = "".join(
             [
                 API_URL + "/dataset/",
@@ -566,10 +575,23 @@ def single_upload(
             annotation_response = self.__annotation_upload(annotation_path, image_id)
             # Check if upload was a success
             try:
-                annotation_success = annotation_response.json()["success"]
-            except Exception:
-                warnings.warn(f"Bad response: {response}")
+                response_data = annotation_response.json()
+                if "success" in response_data.keys():
+                    annotation_success = True
+                elif "error" in response_data.keys():
+                    warnings.warn(
+                        f"Uploading annotation data for image failed: {str(response_data['error'])}"
+                    )
+                    annotation_success = False
+                else:
+                    warnings.warn(
+                        f"Uploading annotation data for image failed: {str(response_data)}"
+                    )
+                    annotation_success = False
+            except:
+                warnings.warn(f"Bad response: {response.status_code}")
                 annotation_success = False
+
             # Give user warning that annotation failed to upload
             if not annotation_success:
                 warnings.warn(