truefoundry 0.10.7rc1__tar.gz → 0.11.1rc1__tar.gz

{truefoundry-0.10.7rc1 → truefoundry-0.11.1rc1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: truefoundry
-Version: 0.10.7rc1
+Version: 0.11.1rc1
 Summary: TrueFoundry CLI
 Author-email: TrueFoundry Team <abhishek@truefoundry.com>
 Requires-Python: <3.14,>=3.8.1
@@ -30,7 +30,7 @@ Requires-Dist: requirements-parser<0.12.0,>=0.11.0
 Requires-Dist: rich-click<2.0.0,>=1.2.1
 Requires-Dist: rich<14.0.0,>=13.7.1
 Requires-Dist: tqdm<5.0.0,>=4.0.0
-Requires-Dist: truefoundry-sdk<0.2.0,>=0.1.3
+Requires-Dist: truefoundry-sdk<0.2.0,>=0.1.7
 Requires-Dist: typing-extensions>=4.0
 Requires-Dist: urllib3<3,>=1.26.18
 Requires-Dist: yq<4.0.0,>=3.1.0

{truefoundry-0.10.7rc1 → truefoundry-0.11.1rc1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "truefoundry"
-version = "0.10.7rc1"  # This is a placeholder version, it will be replaced when releasing
+version = "0.11.1rc1"  # This is a placeholder version, it will be replaced when releasing
 requires-python = ">=3.8.1,<3.14"
 description = "TrueFoundry CLI"
 authors = [{ name = "TrueFoundry Team", email = "abhishek@truefoundry.com" }]
@@ -31,7 +31,7 @@ dependencies = [
   "rich-click>=1.2.1,<2.0.0",
   "rich>=13.7.1,<14.0.0",
   "tqdm>=4.0.0,<5.0.0",
-  "truefoundry-sdk>=0.1.3,<0.2.0",
+  "truefoundry-sdk>=0.1.7,<0.2.0",
   "typing-extensions>=4.0",
   "urllib3>=1.26.18,<3",
   "yq>=3.1.0,<4.0.0",

{truefoundry-0.10.7rc1 → truefoundry-0.11.1rc1}/truefoundry/deploy/builder/builders/tfy_spark_buildpack/dockerfile_template.py

@@ -47,7 +47,12 @@ RUN apt update && \
     + _POST_USER_TEMPLATE
 )
 
-ADDITIONAL_PIP_PACKAGES = ["papermill>=2.6.0,<2.7.0", "ipykernel>=6.0.0,<7.0.0"]
+ADDITIONAL_PIP_PACKAGES = [
+    "papermill>=2.6.0,<2.7.0",
+    "ipykernel>=6.0.0,<7.0.0",
+    "nbconvert>=7.16.6,<7.17.0",
+    "boto3>=1.38.43,<1.40.0",
+]
 
 
 def generate_pip_install_command(

truefoundry-0.11.1rc1/truefoundry/deploy/builder/builders/tfy_spark_buildpack/tfy_execute_notebook.py

@@ -0,0 +1,154 @@
+# This script is used as the main application file for spark applications
+# when the application to be run is a notebook, the actual notebook to be
+# executed is passed as an argument to this script.
+
+
+import argparse
+import os
+import sys
+
+import boto3
+import nbformat
+import papermill as pm
+from botocore.client import Config
+from nbconvert import HTMLExporter
+
+
+def convert_notebook_to_html(notebook_path, output_html_path):
+    """
+    Convert a Jupyter notebook to an HTML file.
+
+    Args:
+        notebook_path: Path to the input notebook (.ipynb)
+        output_html_path: Path for the output HTML file (.html)
+    """
+    print(f"Converting notebook {notebook_path} to HTML...")
+    try:
+        with open(notebook_path, "r", encoding="utf-8") as f:
+            notebook_content = nbformat.read(f, as_version=4)
+
+        html_exporter = HTMLExporter()
+        # Use lab for https://nbconvert.readthedocs.io/en/latest/customizing.html#where-are-nbconvert-templates-installed
+        html_exporter.template_name = "lab"
+        (body, _) = html_exporter.from_notebook_node(notebook_content)
+
+        with open(output_html_path, "w", encoding="utf-8") as f:
+            f.write(body)
+        print(f"Successfully converted notebook to HTML: {output_html_path}")
+    except Exception as e:
+        print(f"Error converting notebook to HTML: {e}")
+        raise
+
+
+def upload_file_to_s3(file_path, bucket_name, s3_key):
+    print(f"Uploading {file_path} to s3://{bucket_name}/{s3_key}...")
+    # Use s3proxy for pushing data to s3
+    # The JWT token is already available in the pod
+    aws_access_key_id = os.environ.get("SPARK_APPLICATION_EVENT_LOG_JWT_TOKEN")
+    aws_secret_access_key = "__token__"
+    s3_endpoint_url = os.environ.get("S3_PROXY_URL")
+
+    if not aws_access_key_id:
+        raise ValueError(
+            "SPARK_APPLICATION_EVENT_LOG_JWT_TOKEN environment variable is not set"
+        )
+    if not s3_endpoint_url:
+        raise ValueError("S3_PROXY_URL environment variable is not set")
+
+    # Needed for the issue https://github.com/gaul/s3proxy/issues/765
+    s3_config = Config(
+        request_checksum_calculation="when_required",
+        response_checksum_validation="when_required",
+    )
+    try:
+        client = boto3.client(
+            "s3",
+            aws_access_key_id=aws_access_key_id,
+            aws_secret_access_key=aws_secret_access_key,
+            endpoint_url=s3_endpoint_url,
+            config=s3_config,
+        )
+
+        with open(file_path, "rb") as data:
+            client.put_object(Bucket=bucket_name, Key=s3_key, Body=data)
+        print(f"Successfully uploaded {file_path} to s3://{bucket_name}/{s3_key}")
+    except Exception as e:
+        print(f"Error uploading file to S3: {e}")
+        raise
+
+
+def execute_notebook(notebook_path, output_path="/tmp/output.ipynb", parameters=None):
+    """
+    Execute a Jupyter notebook using papermill.
+
+    Args:
+        notebook_path: Path to the input notebook
+        output_path: Path for the output notebook
+        parameters: Dictionary of parameters to pass to the notebook
+
+    Raises:
+        Exception: If notebook execution fails
+    """
+    if parameters is None:
+        parameters = {}
+
+    print(f"Starting execution of notebook: {notebook_path}")
+    pm.execute_notebook(
+        input_path=notebook_path,
+        output_path=output_path,
+        parameters=parameters,
+        # TODO(gw): Replace with kernel name for venv
+        kernel_name="python3",
+        # Log cell by cell execution output
+        # TODO(gw): Output logs to a file instead, so that they aren't merged with the container's logs
+        log_output=True,
+        stdout_file=sys.stdout,
+        stderr_file=sys.stderr,
+    )
+    print(f"Successfully executed notebook: {notebook_path}")
+    return output_path
+
+
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(
+        description="Execute a Jupyter notebook using papermill for Spark applications"
+    )
+    parser.add_argument("notebook_path", help="Path to the notebook file to execute")
+    args = parser.parse_args()
+
+    output_notebook_path = "/tmp/output.ipynb"
+
+    # This would be the same as the default bucket used by servicefoundry-server
+    s3_bucket = os.environ.get("TFY_NOTEBOOK_OUTPUT_S3_BUCKET")
+    # This would be something like sparkjob-events/<tenant-id>
+    s3_key_prefix = os.environ.get("TFY_NOTEBOOK_OUTPUT_S3_KEY_PREFIX")
+
+    try:
+        executed_notebook_path = execute_notebook(
+            args.notebook_path, output_path=output_notebook_path
+        )
+
+        # The following may also be modeled as an entrypoint
+        # https://papermill.readthedocs.io/en/latest/extending-entry-points.html
+        # Will take that up with next iteration where we save the executed notebook periodically
+        if s3_bucket and s3_key_prefix:
+            print("Converting notebook to HTML and uploading to S3...")
+            html_output_path = "/tmp/output.html"
+            convert_notebook_to_html(
+                notebook_path=executed_notebook_path, output_html_path=html_output_path
+            )
+
+            # Construct S3 key: use the original notebook name for the HTML file
+            notebook_name = os.path.basename(args.notebook_path)
+            s3_html_key = f"{s3_key_prefix}/output.html"
+            upload_file_to_s3(
+                file_path=html_output_path, bucket_name=s3_bucket, s3_key=s3_html_key
+            )
+            print(f"Successfully uploaded HTML to s3://{s3_bucket}/{s3_html_key}")
+
+    except Exception as e:
+        print(f"Error executing notebook {args.notebook_path}: {e}")
+        print(
+            "Exiting with status code 1 to signal failure to parent process/orchestrator"
+        )
+        sys.exit(1)

{truefoundry-0.10.7rc1 → truefoundry-0.11.1rc1}/truefoundry/ml/artifact/truefoundry_artifact_repo.py

@@ -30,7 +30,14 @@ from rich.progress import (
     TransferSpeedColumn,
 )
 from tqdm.utils import CallbackIOWrapper
+from truefoundry_sdk import (
+    MultiPartUploadResponse,
+    MultiPartUploadStorageProvider,
+    Operation,
+    SignedUrl,
+)
 
+from truefoundry import client
 from truefoundry.common.constants import ENV_VARS
 from truefoundry.common.request_utils import (
     augmented_raise_for_status,
@@ -45,23 +52,13 @@ from truefoundry.common.storage_provider_utils import (
 )
 from truefoundry.ml._autogen.client import (  # type: ignore[attr-defined]
     ApiClient,
-    CreateMultiPartUploadForDatasetRequestDto,
-    CreateMultiPartUploadRequestDto,
     FileInfoDto,
-    GetSignedURLForDatasetWriteRequestDto,
-    GetSignedURLsForArtifactVersionReadRequestDto,
-    GetSignedURLsForArtifactVersionWriteRequestDto,
-    GetSignedURLsForDatasetReadRequestDto,
     ListFilesForArtifactVersionRequestDto,
     ListFilesForArtifactVersionsResponseDto,
     ListFilesForDatasetRequestDto,
     ListFilesForDatasetResponseDto,
     MlfoundryArtifactsApi,
-    MultiPartUploadDto,
-    MultiPartUploadResponseDto,
-    MultiPartUploadStorageProvider,
     RunArtifactsApi,
-    SignedURLDto,
 )
 from truefoundry.ml.exceptions import MlFoundryException
 from truefoundry.ml.logger import logger
@@ -127,7 +124,7 @@ def verify_artifact_path(artifact_path):
 
 
 def _signed_url_upload_file(
-    signed_url: SignedURLDto,
+    signed_url: SignedUrl,
     local_file: str,
     progress_bar: Progress,
     abort_event: Optional[Event] = None,
@@ -206,21 +203,20 @@ def _any_future_has_failed(futures) -> bool:
 
 class ArtifactIdentifier(BaseModel):
     artifact_version_id: Optional[uuid.UUID] = None
+    dataset_id: Optional[str] = None
     dataset_fqn: Optional[str] = None
 
     @root_validator
     def _check_identifier_type(cls, values: Dict[str, Any]):
         if not values.get("artifact_version_id", False) and not values.get(
-            "dataset_fqn", False
+            "dataset_id", False
         ):
             raise MlFoundryException(
-                "One of the version_id or dataset_fqn should be passed"
+                "One of the version_id or dataset_id should be passed"
             )
-        if values.get("artifact_version_id", False) and values.get(
-            "dataset_fqn", False
-        ):
+        if values.get("artifact_version_id", False) and values.get("dataset_id", False):
             raise MlFoundryException(
-                "Exactly one of version_id or dataset_fqn should be passed"
+                "Exactly one of version_id or dataset_id should be passed"
             )
         return values
 
@@ -243,24 +239,25 @@ class MlFoundryArtifactsRepository:
         self,
         artifact_identifier: ArtifactIdentifier,
         paths,
-    ) -> List[SignedURLDto]:
+    ) -> List[SignedUrl]:
         if artifact_identifier.artifact_version_id:
-            signed_urls_response = self._mlfoundry_artifacts_api.get_signed_urls_for_read_post(
-                get_signed_urls_for_artifact_version_read_request_dto=GetSignedURLsForArtifactVersionReadRequestDto(
-                    id=str(artifact_identifier.artifact_version_id), paths=paths
-                )
+            signed_urls_response = client.artifact_versions.get_signed_urls(
+                id=str(artifact_identifier.artifact_version_id),
+                paths=paths,
+                operation=Operation.READ,
             )
-            signed_urls = signed_urls_response.signed_urls
-        elif artifact_identifier.dataset_fqn:
-            signed_urls_dataset_response = self._mlfoundry_artifacts_api.get_signed_urls_dataset_read_post(
-                get_signed_urls_for_dataset_read_request_dto=GetSignedURLsForDatasetReadRequestDto(
-                    dataset_fqn=artifact_identifier.dataset_fqn, paths=paths
-                )
+            signed_urls = signed_urls_response.data
+        elif artifact_identifier.dataset_id:
+            signed_urls_dataset_response = client.data_directories.get_signed_urls(
+                id=str(artifact_identifier.dataset_id),
+                paths=paths,
+                operation=Operation.READ,
             )
-            signed_urls = signed_urls_dataset_response.signed_urls
+            signed_urls = signed_urls_dataset_response.data
+
         else:
             raise ValueError(
-                "Invalid artifact type - both `artifact_version_id` and `dataset_fqn` both are None"
+                "Invalid artifact type - both `artifact_version_id` and `dataset_id` both are None"
             )
         return signed_urls
 
@@ -268,24 +265,24 @@ class MlFoundryArtifactsRepository:
         self,
         artifact_identifier: ArtifactIdentifier,
         paths: List[str],
-    ) -> List[SignedURLDto]:
+    ) -> List[SignedUrl]:
         if artifact_identifier.artifact_version_id:
-            signed_urls_response = self._mlfoundry_artifacts_api.get_signed_urls_for_write_post(
-                get_signed_urls_for_artifact_version_write_request_dto=GetSignedURLsForArtifactVersionWriteRequestDto(
-                    id=str(artifact_identifier.artifact_version_id), paths=paths
-                )
+            signed_urls_response = client.artifact_versions.get_signed_urls(
+                id=str(artifact_identifier.artifact_version_id),
+                paths=paths,
+                operation=Operation.WRITE,
             )
-            signed_urls = signed_urls_response.signed_urls
-        elif artifact_identifier.dataset_fqn:
-            signed_urls_dataset_response = self._mlfoundry_artifacts_api.get_signed_urls_for_dataset_write_post(
-                get_signed_url_for_dataset_write_request_dto=GetSignedURLForDatasetWriteRequestDto(
-                    dataset_fqn=artifact_identifier.dataset_fqn, paths=paths
-                )
+            signed_urls = signed_urls_response.data
+        elif artifact_identifier.dataset_id:
+            signed_urls_dataset_response = client.data_directories.get_signed_urls(
+                id=str(artifact_identifier.dataset_id),
+                paths=paths,
+                operation=Operation.WRITE,
             )
-            signed_urls = signed_urls_dataset_response.signed_urls
+            signed_urls = signed_urls_dataset_response.data
         else:
             raise ValueError(
-                "Invalid artifact type - both `artifact_version_id` and `dataset_fqn` both are None"
+                "Invalid artifact type - both `artifact_version_id` and `dataset_id` both are None"
             )
         return signed_urls
 
@@ -293,7 +290,7 @@ class MlFoundryArtifactsRepository:
         self,
         local_file: str,
         artifact_path: str,
-        signed_url: Optional[SignedURLDto],
+        signed_url: Optional[SignedUrl],
         progress_bar: Progress,
         abort_event: Optional[Event] = None,
     ):
@@ -322,28 +319,30 @@ class MlFoundryArtifactsRepository:
         artifact_identifier: ArtifactIdentifier,
         path,
         num_parts,
-    ) -> MultiPartUploadDto:
+    ) -> MultiPartUpload:
         if artifact_identifier.artifact_version_id:
-            create_multipart_response: MultiPartUploadResponseDto = self._mlfoundry_artifacts_api.create_multi_part_upload_post(
-                create_multi_part_upload_request_dto=CreateMultiPartUploadRequestDto(
-                    artifact_version_id=str(artifact_identifier.artifact_version_id),
+            create_multipart_response: MultiPartUploadResponse = (
+                client.artifact_versions.create_multi_part_upload(
+                    id=str(artifact_identifier.artifact_version_id),
                     path=path,
                     num_parts=num_parts,
                 )
             )
-            multipart_upload = create_multipart_response.multipart_upload
-        elif artifact_identifier.dataset_fqn:
-            create_multipart_for_dataset_response = self._mlfoundry_artifacts_api.create_multipart_upload_for_dataset_post(
-                create_multi_part_upload_for_dataset_request_dto=CreateMultiPartUploadForDatasetRequestDto(
-                    dataset_fqn=artifact_identifier.dataset_fqn,
+            multipart_upload = create_multipart_response.data
+
+        elif artifact_identifier.dataset_id:
+            create_multipart_for_dataset_response: MultiPartUploadResponse = (
+                client.data_directories.create_multipart_upload(
+                    id=str(artifact_identifier.dataset_id),
                     path=path,
                     num_parts=num_parts,
                 )
             )
-            multipart_upload = create_multipart_for_dataset_response.multipart_upload
+            multipart_upload = create_multipart_for_dataset_response.data
+
         else:
             raise ValueError(
-                "Invalid artifact type - both `artifact_version_id` and `dataset_fqn` both are None"
+                "Invalid artifact type - both `artifact_version_id` and `dataset_id` both are None"
             )
         return multipart_upload
 
@@ -370,10 +369,10 @@ class MlFoundryArtifactsRepository:
         )
         if (
             multipart_upload.storage_provider
-            is MultiPartUploadStorageProvider.S3_COMPATIBLE
+            is MultiPartUploadStorageProvider.S3COMPATIBLE
         ):
             s3_compatible_multipart_upload(
-                multipart_upload=MultiPartUpload.parse_obj(multipart_upload.to_dict()),
+                multipart_upload=MultiPartUpload.parse_obj(multipart_upload.dict()),
                 local_file=local_file,
                 executor=executor,
                 multipart_info=multipart_info,
@@ -403,7 +402,7 @@ class MlFoundryArtifactsRepository:
         artifact_path: str,
         multipart_info: _FileMultiPartInfo,
         progress_bar: Progress,
-        signed_url: Optional[SignedURLDto] = None,
+        signed_url: Optional[SignedUrl] = None,
         abort_event: Optional[Event] = None,
         executor_for_multipart_upload: Optional[ThreadPoolExecutor] = None,
     ):
@@ -678,7 +677,7 @@ class MlFoundryArtifactsRepository:
         remote_file_path: str,
         local_path: str,
         progress_bar: Optional[Progress],
-        signed_url: Optional[SignedURLDto],
+        signed_url: Optional[SignedUrl],
         abort_event: Optional[Event] = None,
     ):
         if not remote_file_path:
@@ -719,7 +718,7 @@ class MlFoundryArtifactsRepository:
         self,
         src_artifact_path: str,
         dst_local_dir_path: str,
-        signed_url: Optional[SignedURLDto],
+        signed_url: Optional[SignedUrl],
         progress_bar: Optional[Progress] = None,
         abort_event=None,
     ) -> str:

{truefoundry-0.10.7rc1 → truefoundry-0.11.1rc1}/truefoundry/ml/log_types/artifacts/artifact.py

@@ -8,6 +8,7 @@ import warnings
 from pathlib import Path
 from typing import TYPE_CHECKING, Any, Dict, NamedTuple, Optional, Union
 
+from truefoundry import client
 from truefoundry.common.warnings import TrueFoundryDeprecationWarning
 from truefoundry.ml._autogen.client import (  # type: ignore[attr-defined]
     ArtifactDto,
@@ -16,7 +17,6 @@ from truefoundry.ml._autogen.client import (  # type: ignore[attr-defined]
     ArtifactVersionDto,
     CreateArtifactVersionRequestDto,
     DeleteArtifactVersionsRequestDto,
-    ExperimentsApi,
     ExternalBlobStorageSource,
     FinalizeArtifactVersionRequestDto,
     MlfoundryArtifactsApi,
@@ -444,19 +444,15 @@ def _log_artifact_version_helper(
 
     if run:
         mlfoundry_artifacts_api = run._mlfoundry_artifacts_api
-        experiments_api = run._experiments_api
-        ml_repo = experiments_api.get_experiment_get(
-            experiment_id=str(run._experiment_id)
-        ).experiment.name
+        repos = client.ml_repos.get(id=run._experiment_id)
+        ml_repo = repos.data.manifest.name
         ml_repo_id = run._experiment_id
     else:
         assert ml_repo is not None
         api_client = _get_api_client()
         mlfoundry_artifacts_api = MlfoundryArtifactsApi(api_client=api_client)
-        experiments_api = ExperimentsApi(api_client=api_client)
-        ml_repo_id = experiments_api.get_experiment_by_name_get(
-            experiment_name=ml_repo
-        ).experiment.experiment_id
+        result = list(client.ml_repos.list(name=ml_repo, limit=1))[0]
+        ml_repo_id = result.id
 
     metadata = metadata or {}
     if ARTIFACT_METADATA_TRUEFOUNDRY_KEY not in metadata:

{truefoundry-0.10.7rc1 → truefoundry-0.11.1rc1}/truefoundry/ml/log_types/artifacts/dataset.py

@@ -84,7 +84,9 @@ class DataDirectory:
 
     def _get_artifacts_repo(self):
         return MlFoundryArtifactsRepository(
-            artifact_identifier=ArtifactIdentifier(dataset_fqn=self._dataset.fqn),
+            artifact_identifier=ArtifactIdentifier(
+                dataset_id=self._dataset.id, dataset_fqn=self._dataset.fqn
+            ),
             api_client=self._api_client,
         )
 

{truefoundry-0.10.7rc1 → truefoundry-0.11.1rc1}/truefoundry/ml/log_types/artifacts/model.py

@@ -9,12 +9,12 @@ import warnings
 from pathlib import Path
 from typing import TYPE_CHECKING, Any, Dict, Optional, Union
 
+from truefoundry import client
 from truefoundry.common.warnings import TrueFoundryDeprecationWarning
 from truefoundry.ml._autogen.client import (  # type: ignore[attr-defined]
     ArtifactType,
     CreateArtifactVersionRequestDto,
     DeleteArtifactVersionsRequestDto,
-    ExperimentsApi,
     ExternalBlobStorageSource,
     FinalizeArtifactVersionRequestDto,
     Framework,
@@ -531,19 +531,15 @@ def _log_model_version(  # noqa: C901
 
     if run:
         mlfoundry_artifacts_api = run._mlfoundry_artifacts_api
-        experiments_api = run._experiments_api
-        ml_repo = experiments_api.get_experiment_get(
-            experiment_id=str(run._experiment_id)
-        ).experiment.name
+        repos = client.ml_repos.get(id=run._experiment_id)
+        ml_repo = repos.data.manifest.name
         ml_repo_id = run._experiment_id
     else:
         assert ml_repo is not None
         api_client = _get_api_client()
         mlfoundry_artifacts_api = MlfoundryArtifactsApi(api_client=api_client)
-        experiments_api = ExperimentsApi(api_client=api_client)
-        ml_repo_id = experiments_api.get_experiment_by_name_get(
-            experiment_name=ml_repo
-        ).experiment.experiment_id
+        result = list(client.ml_repos.list(name=ml_repo, limit=1))[0]
+        ml_repo_id = result.id
 
     step = step or 0
     total_size = None