api tagging for pipeline/train/evaluate

modelscope/hub/api.py

@@ -646,6 +646,30 @@ class HubApi:
     def check_local_cookies(self, use_cookies) -> CookieJar:
         return self._check_cookie(use_cookies=use_cookies)
 
+    def create_library_statistics(self,
+                                  method: str,
+                                  name: str,
+                                  cn_name: Optional[str]):
+        """
+            create library statistics. called by train()/evaluate()/pipeline()
+
+        Args:
+            method (str): called methed name,i.e train/evaluate/pipeline
+            name (str): model name, for example: damo/cv_unet_person-image-cartoon_compound-models
+            cn_name (str): model name in chinese, for example: 达摩卡通化模型
+        Raises:
+            ValueError: If user_cookies is True, but no local cookie.
+
+        Returns:
+            None
+        """
+        path = f'{self.endpoint}/api/v1/statistics/library'
+        headers = {'user-agent': ModelScopeConfig.get_user_agent()}
+        params = {"Method": method, "Name": name, "CnName": cn_name}
+        r = requests.post(path, params=params, headers=headers)
+        r.raise_for_status()
+        return
+
 
 class ModelScopeConfig:
     path_credential = expanduser(DEFAULT_CREDENTIALS_PATH)

modelscope/pipelines/base.py

@@ -23,6 +23,7 @@ from modelscope.utils.hub import read_config, snapshot_download
 from modelscope.utils.import_utils import is_tf_available, is_torch_available
 from modelscope.utils.logger import get_logger
 from modelscope.utils.torch_utils import _find_free_port, _is_free_port
+from modelscope.hub.api import HubApi
 from .util import is_model, is_official_hub_path
 
 if is_torch_available():
@@ -151,7 +152,9 @@ class Pipeline(ABC):
                  **kwargs) -> Union[Dict[str, Any], Generator]:
         # model provider should leave it as it is
         # modelscope library developer will handle this function
-
+        _api = HubApi()
+        model_name = self.cfg.task
+        _api.create_library_statistics("pipeline", model_name, None)
         # place model to cpu or gpu
         if (self.model or (self.has_multiple_models and self.models[0])):
             if not self._model_prepare:

modelscope/trainers/trainer.py

@@ -39,6 +39,7 @@ from modelscope.utils.logger import get_logger
 from modelscope.utils.registry import build_from_cfg
 from modelscope.utils.torch_utils import (get_dist_info, get_local_rank,
                                           init_dist, set_random_seed)
+from modelscope.hub.api import HubApi
 from .base import BaseTrainer
 from .builder import TRAINERS
 from .default_config import merge_cfg
@@ -436,6 +437,9 @@ class EpochBasedTrainer(BaseTrainer):
 
     def train(self, checkpoint_path=None, *args, **kwargs):
         self._mode = ModeKeys.TRAIN
+        _api = HubApi()
+        model_name = self.cfg.task
+        _api.create_library_statistics("train", model_name, None)
 
         if self.train_dataset is None:
             self.train_dataloader = self.get_train_dataloader()
@@ -456,6 +460,9 @@ class EpochBasedTrainer(BaseTrainer):
         self.train_loop(self.train_dataloader)
 
     def evaluate(self, checkpoint_path=None):
+        _api = HubApi()
+        model_name = self.cfg.task
+        _api.create_library_statistics("evaluate", model_name, None)
         if checkpoint_path is not None and os.path.isfile(checkpoint_path):
             from modelscope.trainers.hooks import CheckpointHook
             CheckpointHook.load_checkpoint(checkpoint_path, self)