futurehouse-client 0.3.17.dev94__tar.gz → 0.3.18.dev80__tar.gz

{futurehouse_client-0.3.17.dev94 → futurehouse_client-0.3.18.dev80}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: futurehouse-client
-Version: 0.3.17.dev94
+Version: 0.3.18.dev80
 Summary: A client for interacting with endpoints of the FutureHouse service.
 Author-email: FutureHouse technical staff <hello@futurehouse.org>
 Classifier: Operating System :: OS Independent

{futurehouse_client-0.3.17.dev94 → futurehouse_client-0.3.18.dev80}/futurehouse_client/clients/rest_client.py

@@ -8,6 +8,7 @@ import inspect
 import json
 import logging
 import os
+import sys
 import tempfile
 import time
 import uuid
@@ -63,24 +64,14 @@ from futurehouse_client.utils.monitoring import (
 )
 
 logger = logging.getLogger(__name__)
-
+logging.basicConfig(
+    level=logging.INFO,
+    format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
+    stream=sys.stdout,
+)
+logging.getLogger("httpx").setLevel(logging.WARNING)
 TaskRequest.model_rebuild()
 
-retry_if_connection_error = retry_if_exception_type((
-    # From requests
-    Timeout,
-    ConnectionError,
-    RequestException,
-    # From httpx
-    ConnectError,
-    ConnectTimeout,
-    ReadTimeout,
-    ReadError,
-    NetworkError,
-    RemoteProtocolError,
-    CloseError,
-))
-
 FILE_UPLOAD_IGNORE_PARTS = {
     ".ruff_cache",
     "__pycache__",
@@ -111,6 +102,27 @@ class InvalidTaskDescriptionError(Exception):
     """Raised when the task description is invalid or empty."""
 
 
+class FileUploadError(RestClientError):
+    """Raised when there's an error uploading a file."""
+
+
+retry_if_connection_error = retry_if_exception_type((
+    # From requests
+    Timeout,
+    ConnectionError,
+    RequestException,
+    # From httpx
+    ConnectError,
+    ConnectTimeout,
+    ReadTimeout,
+    ReadError,
+    NetworkError,
+    RemoteProtocolError,
+    CloseError,
+    FileUploadError,
+))
+
+
 class SimpleOrganization(BaseModel):
     id: int
     name: str
@@ -207,10 +219,6 @@ class TaskResponseVerbose(TaskResponse):
     shared_with: list[SimpleOrganization] | None = None
 
 
-class FileUploadError(RestClientError):
-    """Raised when there's an error uploading a file."""
-
-
 class RestClient:
     REQUEST_TIMEOUT: ClassVar[float] = 30.0  # sec
     MAX_RETRY_ATTEMPTS: ClassVar[int] = 3
@@ -917,14 +925,14 @@ class RestClient:
         self,
         job_name: str,
         file_path: str | os.PathLike,
-        folder_name: str | None = None,
+        upload_id: str | None = None,
     ) -> str:
         """Upload a file or directory to a futurehouse job bucket.
 
         Args:
             job_name: The name of the futurehouse job to upload to.
             file_path: The local path to the file or directory to upload.
-            folder_name: Optional folder name to use for the upload. If not provided, a random UUID will be used.
+            upload_id: Optional folder name to use for the upload. If not provided, a random UUID will be used.
 
         Returns:
             The upload ID used for the upload.
@@ -936,7 +944,7 @@ class RestClient:
         if not file_path.exists():
             raise FileNotFoundError(f"File or directory not found: {file_path}")
 
-        upload_id = folder_name or str(uuid.uuid4())
+        upload_id = upload_id or str(uuid.uuid4())
 
         if file_path.is_dir():
             # Process directory recursively
@@ -999,6 +1007,12 @@ class RestClient:
         """
         file_name = file_name or file_path.name
         file_size = file_path.stat().st_size
+
+        # Skip empty files
+        if file_size == 0:
+            logger.warning(f"Skipping upload of empty file: {file_path}")
+            return
+
         total_chunks = (file_size + self.CHUNK_SIZE - 1) // self.CHUNK_SIZE
 
         logger.info(f"Uploading {file_path} as {file_name} ({total_chunks} chunks)")
@@ -1056,12 +1070,18 @@ class RestClient:
         wait=wait_exponential(multiplier=RETRY_MULTIPLIER, max=MAX_RETRY_WAIT),
         retry=retry_if_connection_error,
     )
-    def list_files(self, job_name: str, folder_name: str) -> dict[str, list[str]]:
+    def list_files(
+        self,
+        job_name: str,
+        trajectory_id: str | None = None,
+        upload_id: str | None = None,
+    ) -> dict[str, list[str]]:
         """List files and directories in a GCS location for a given job_name and upload_id.
 
         Args:
             job_name: The name of the futurehouse job.
-            folder_name: The specific folder name (upload_id) to list files from.
+            trajectory_id: The specific trajectory id to list files from.
+            upload_id: The specific upload id to list files from.
 
         Returns:
             A list of files in the GCS folder.
@@ -1069,22 +1089,27 @@ class RestClient:
         Raises:
             RestClientError: If there is an error listing the files.
         """
+        if not bool(trajectory_id) ^ bool(upload_id):
+            raise RestClientError(
+                "Must at least specify one of trajectory_id or upload_id, but not both"
+            )
         try:
             url = f"/v0.1/crows/{job_name}/list-files"
-            params = {"upload_id": folder_name}
+            params = {"trajectory_id": trajectory_id, "upload_id": upload_id}
+            params = {k: v for k, v in params.items() if v is not None}
             response = self.client.get(url, params=params)
             response.raise_for_status()
             return response.json()
         except HTTPStatusError as e:
             logger.exception(
-                f"Error listing files for job {job_name}, folder {folder_name}: {e.response.text}"
+                f"Error listing files for job {job_name}, trajectory {trajectory_id}, upload_id {upload_id}: {e.response.text}"
             )
             raise RestClientError(
                 f"Error listing files: {e.response.status_code} - {e.response.text}"
             ) from e
         except Exception as e:
             logger.exception(
-                f"Error listing files for job {job_name}, folder {folder_name}"
+                f"Error listing files for job {job_name}, trajectory {trajectory_id}, upload_id {upload_id}"
             )
             raise RestClientError(f"Error listing files: {e!s}") from e
 
@@ -1096,7 +1121,7 @@ class RestClient:
     def download_file(
         self,
         job_name: str,
-        folder_name: str,
+        trajectory_id: str,
         file_path: str,
         destination_path: str | os.PathLike,
     ) -> None:
@@ -1104,14 +1129,14 @@ class RestClient:
 
         Args:
             job_name: The name of the futurehouse job.
-            folder_name: The specific folder name (upload_id) the file belongs to.
+            trajectory_id: The specific trajectory id the file belongs to.
             file_path: The relative path of the file to download
                        (e.g., 'data/my_file.csv' or 'my_image.png').
             destination_path: The local path where the file should be saved.
 
         Raises:
             RestClientError: If there is an error downloading the file.
-            FileNotFoundError: If the destination directory does not exist.
+            FileNotFoundError: If the destination directory does not exist or if the file is not found.
         """
         destination_path = Path(destination_path)
         # Ensure the destination directory exists
@@ -1119,17 +1144,24 @@ class RestClient:
 
         try:
             url = f"/v0.1/crows/{job_name}/download-file"
-            params = {"upload_id": folder_name, "file_path": file_path}
+            params = {"trajectory_id": trajectory_id, "file_path": file_path}
 
             with self.client.stream("GET", url, params=params) as response:
                 response.raise_for_status()  # Check for HTTP errors before streaming
                 with open(destination_path, "wb") as f:
                     for chunk in response.iter_bytes(chunk_size=8192):
                         f.write(chunk)
+
+            # Check if the downloaded file is empty
+            if destination_path.stat().st_size == 0:
+                # Remove the empty file
+                destination_path.unlink()
+                raise FileNotFoundError(f"File not found or is empty: {file_path}")
+
             logger.info(f"File {file_path} downloaded to {destination_path}")
         except HTTPStatusError as e:
             logger.exception(
-                f"Error downloading file {file_path} for job {job_name}, folder {folder_name}: {e.response.text}"
+                f"Error downloading file {file_path} for job {job_name}, trajectory_id {trajectory_id}: {e.response.text}"
             )
             # Clean up partially downloaded file if an error occurs
             if destination_path.exists():
@@ -1137,9 +1169,20 @@ class RestClient:
             raise RestClientError(
                 f"Error downloading file: {e.response.status_code} - {e.response.text}"
             ) from e
+        except RemoteProtocolError as e:
+            logger.error(
+                f"Connection error while downloading file {file_path} for job {job_name}, trajectory_id {trajectory_id}"
+            )
+            # Clean up partially downloaded file
+            if destination_path.exists():
+                destination_path.unlink()
+
+            # Often RemoteProtocolError during download means the file wasn't found
+            # or was empty/corrupted on the server side
+            raise FileNotFoundError(f"File not found or corrupted: {file_path}") from e
         except Exception as e:
             logger.exception(
-                f"Error downloading file {file_path} for job {job_name}, folder {folder_name}"
+                f"Error downloading file {file_path} for job {job_name}, trajectory_id {trajectory_id}"
             )
             if destination_path.exists():
                 destination_path.unlink()  # Clean up partial file

{futurehouse_client-0.3.17.dev94 → futurehouse_client-0.3.18.dev80}/futurehouse_client.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: futurehouse-client
-Version: 0.3.17.dev94
+Version: 0.3.18.dev80
 Summary: A client for interacting with endpoints of the FutureHouse service.
 Author-email: FutureHouse technical staff <hello@futurehouse.org>
 Classifier: Operating System :: OS Independent