validmind 2.4.0__tar.gz → 2.4.4__tar.gz

{validmind-2.4.0 → validmind-2.4.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: validmind
-Version: 2.4.0
+Version: 2.4.4
 Summary: ValidMind Developer Framework
 License: Commercial License
 Author: Andres Rodriguez
@@ -46,6 +46,7 @@ Requires-Dist: scikit-learn
 Requires-Dist: scipy
 Requires-Dist: scorecardpy (>=0.1.9.6,<0.2.0.0)
 Requires-Dist: seaborn
+Requires-Dist: sentencepiece (>=0.2.0,<0.3.0) ; extra == "all" or extra == "huggingface" or extra == "llm"
 Requires-Dist: sentry-sdk (>=1.24.0,<2.0.0)
 Requires-Dist: shap (>=0.42.0,<0.43.0)
 Requires-Dist: statsmodels
@@ -53,7 +54,7 @@ Requires-Dist: tabulate (>=0.8.9,<0.9.0)
 Requires-Dist: textblob (>=0.18.0.post0,<0.19.0)
 Requires-Dist: torch (>=1.10.0) ; extra == "all" or extra == "llm" or extra == "pytorch"
 Requires-Dist: tqdm
-Requires-Dist: transformers (>=4.32.0,<5.0.0) ; extra == "all" or extra == "llm" or extra == "huggingface"
+Requires-Dist: transformers (>=4.32.0,<5.0.0) ; extra == "all" or extra == "huggingface" or extra == "llm"
 Requires-Dist: xgboost (>=1.5.2,<3)
 Requires-Dist: ydata-profiling
 Description-Content-Type: text/markdown

{validmind-2.4.0 → validmind-2.4.4}/pyproject.toml

@@ -10,10 +10,9 @@ description = "ValidMind Developer Framework"
 license = "Commercial License"
 name = "validmind"
 readme = "README.pypi.md"
-version = "2.4.0"
+version = "2.4.4"
 
 [tool.poetry.dependencies]
-python = ">=3.8.1,<3.12"
 aiohttp = {extras = ["speedups"], version = "*"}
 arch = "*"
 bert-score = ">=0.3.13"
@@ -21,7 +20,7 @@ catboost = "*"
 evaluate = "*"
 ipywidgets = "*"
 kaleido = ">=0.2.1,!=0.2.1.post1"
-langchain-openai = { version = ">=0.1.8", optional = true }
+langchain-openai = {version = ">=0.1.8", optional = true}
 langdetect = "*"
 latex2mathml = ">=3.77.0"
 llvmlite = {version = "*", python = ">=3.8,<=3.11"}
@@ -36,6 +35,7 @@ plotly = "*"
 plotly-express = "*"
 polars = "*"
 pycocoevalcap = {version = "^1.2", optional = true}
+python = ">=3.8.1,<3.12"
 python-dotenv = "*"
 ragas = {version = ">=0.1.7", optional = true}
 rouge = ">=1"
@@ -44,6 +44,7 @@ scikit-learn = "*"
 scipy = "*"
 scorecardpy = "^0.1.9.6"
 seaborn = "*"
+sentencepiece = {version = "^0.2.0", optional = true}
 sentry-sdk = "^1.24.0"
 shap = "^0.42.0"
 statsmodels = "*"
@@ -80,19 +81,21 @@ all = [
   "transformers",
   "pycocoevalcap",
   "ragas",
+  "sentencepiece",
   "langchain-openai",
 ]
+huggingface = ["transformers", "sentencepiece"]
 llm = [
   "torch",
   "transformers",
   "openai",
   "pycocoevalcap",
   "ragas",
+  "sentencepiece",
   "langchain-openai",
 ]
 pytorch = ["torch"]
 r-support = ["rpy2"]
-huggingface = ["transformers"]
 
 [build-system]
 build-backend = "poetry.core.masonry.api"

validmind-2.4.4/validmind/__version__.py

@@ -0,0 +1 @@
+__version__ = "2.4.4"

{validmind-2.4.0 → validmind-2.4.4}/validmind/api_client.py

@@ -32,8 +32,7 @@ logger = get_logger(__name__)
 _api_key = os.getenv("VM_API_KEY")
 _api_secret = os.getenv("VM_API_SECRET")
 _api_host = os.getenv("VM_API_HOST")
-
-_project = os.getenv("VM_API_PROJECT")
+_model_cuid = os.getenv("VM_API_MODEL")
 _run_cuid = os.getenv("VM_RUN_CUID")
 
 __api_session: aiohttp.ClientSession = None
@@ -56,7 +55,7 @@ def get_api_config() -> Dict[str, Optional[str]]:
         "VM_API_KEY": _api_key,
         "VM_API_SECRET": _api_secret,
         "VM_API_HOST": _api_host,
-        "VM_API_PROJECT": _project,
+        "VM_API_MODEL": _model_cuid,
         "VM_RUN_CUID": _run_cuid,
     }
 
@@ -65,15 +64,15 @@ def get_api_host() -> Optional[str]:
     return _api_host
 
 
-def get_api_project() -> Optional[str]:
-    return _project
+def get_api_model() -> Optional[str]:
+    return _model_cuid
 
 
 def get_api_headers() -> Dict[str, str]:
     return {
         "X-API-KEY": _api_key,
         "X-API-SECRET": _api_secret,
-        "X-PROJECT-CUID": _project,
+        "X-PROJECT-CUID": _model_cuid,
     }
 
 
@@ -82,6 +81,7 @@ def init(
     api_key: Optional[str] = None,
     api_secret: Optional[str] = None,
     api_host: Optional[str] = None,
+    model: Optional[str] = None,
 ):
     """
     Initializes the API client instances and calls the /ping endpoint to ensure
@@ -91,7 +91,8 @@ def init(
     retrieve them from the environment variables `VM_API_KEY` and `VM_API_SECRET`.
 
     Args:
-        project (str): The project CUID
+        project (str, optional): The project CUID. Alias for model. Defaults to None.
+        model (str, optional): The model CUID. Defaults to None.
         api_key (str, optional): The API key. Defaults to None.
         api_secret (str, optional): The API secret. Defaults to None.
         api_host (str, optional): The API host. Defaults to None.
@@ -99,7 +100,7 @@ def init(
     Raises:
         ValueError: If the API key and secret are not provided
     """
-    global _api_key, _api_secret, _api_host, _run_cuid, _project
+    global _api_key, _api_secret, _api_host, _run_cuid, _model_cuid
 
     if api_key == "...":
         # special case to detect when running a notebook with the standard init snippet
@@ -107,9 +108,9 @@ def init(
         # the notebook
         api_host = api_key = api_secret = project = None
 
-    _project = project or os.getenv("VM_API_PROJECT")
+    _model_cuid = project or model or os.getenv("VM_API_MODEL")
 
-    if _project is None:
+    if _model_cuid is None:
         raise MissingProjectIdError()
 
     _api_key = api_key or os.getenv("VM_API_KEY")
@@ -143,7 +144,7 @@ def _get_session() -> aiohttp.ClientSession:
             {
                 "X-API-KEY": _api_key,
                 "X-API-SECRET": _api_secret,
-                "X-PROJECT-CUID": _project,
+                "X-PROJECT-CUID": _model_cuid,
             }
         )
 
@@ -157,7 +158,7 @@ def __ping() -> Dict[str, Any]:
         headers={
             "X-API-KEY": _api_key,
             "X-API-SECRET": _api_secret,
-            "X-PROJECT-CUID": _project,
+            "X-PROJECT-CUID": _model_cuid,
         },
     )
     if r.status_code != 200:
@@ -172,13 +173,14 @@ def __ping() -> Dict[str, Any]:
     if client_config.project is None:
         ack_connected = True
 
+    # TODO: use model object when backend moves away from project
     client_config.project = client_info["project"]
     client_config.documentation_template = client_info.get("documentation_template", {})
     client_config.feature_flags = client_info.get("feature_flags", {})
 
     if ack_connected:
         logger.info(
-            f"Connected to ValidMind. Project: {client_config.project['name']}"
+            f"Connected to ValidMind... Current Model: {client_config.project['name']}"
             f" ({client_config.project['cuid']})"
         )
 
@@ -563,7 +565,7 @@ def start_run() -> str:
         headers={
             "X-API-KEY": _api_key,
             "X-API-SECRET": _api_secret,
-            "X-PROJECT-CUID": _project,
+            "X-PROJECT-CUID": _model_cuid,
         },
     )
 
@@ -584,7 +586,7 @@ def get_ai_key() -> str:
         headers={
             "X-API-KEY": _api_key,
             "X-API-SECRET": _api_secret,
-            "X-PROJECT-CUID": _project,
+            "X-PROJECT-CUID": _model_cuid,
         },
     )
 

{validmind-2.4.0 → validmind-2.4.4}/validmind/datasets/nlp/cnn_dailymail.py

@@ -15,6 +15,7 @@ text_column = "article"
 target_column = "highlights"
 gpt_35_prediction_column = "gpt_35_prediction"
 t5_prediction = "t5_prediction"
+bert_embedding_prediction_column = "bert_embedding_model_prediction"
 
 # Define the path to the dataset directory
 current_path = os.path.dirname(os.path.abspath(__file__))
@@ -55,7 +56,15 @@ def load_data(source="online", dataset_size=None):
 
         # Load the dataset
         df = pd.read_csv(data_file)
-        df = df[["article", "highlights", "gpt_35_prediction", "t5_prediction"]]
+        df = df[
+            [
+                "article",
+                "highlights",
+                "gpt_35_prediction",
+                "t5_prediction",
+                "bert_embedding_model_prediction",
+            ]
+        ]
 
         train_df = df.sample(frac=0.7, random_state=42)
         test_df = df.drop(train_df.index)