scientiflow-cli 0.4.13__tar.gz → 0.4.16__tar.gz

{scientiflow_cli-0.4.13 → scientiflow_cli-0.4.16}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: scientiflow-cli
-Version: 0.4.13
+Version: 0.4.16
 Summary: CLI tool for scientiflow. This application runs on the client side, decodes pipelines, and executes them in the configured order!
 License: Proprietary
 Author: ScientiFlow

{scientiflow_cli-0.4.13 → scientiflow_cli-0.4.16}/pyproject.toml

@@ -3,7 +3,7 @@ mode = "prod"
 
 [tool.poetry]
 name = "scientiflow-cli"
-version = "0.4.13"
+version = "0.4.16"
 description = "CLI tool for scientiflow. This application runs on the client side, decodes pipelines, and executes them in the configured order!"
 authors = ["ScientiFlow <scientiflow@gmail.com>"]
 license = "Proprietary"

{scientiflow_cli-0.4.13 → scientiflow_cli-0.4.16}/scientiflow_cli/pipeline/decode_and_execute.py

@@ -116,7 +116,7 @@ def execute_background_command_standalone(command: str, log_file_path: str):
         return False
 
 class PipelineExecutor:
-    def __init__(self, base_dir: str, project_id: int, project_job_id: int, project_title: str, job_dir_name: str, nodes: List[Dict[str, Any]], edges: List[Dict[str, str]], environment_variables: Dict[str, str], start_node: str = None, end_node: str = None, job_status: str = None, current_node_from_config: str = None):
+    def __init__(self, base_dir: str, project_id: int, project_job_id: int, project_title: str, job_dir_name: str, nodes: List[Dict[str, Any]], edges: List[Dict[str, str]], environment_variables: Dict[str, str], start_node: str = None, end_node: str = None, job_status: str = None, current_node_from_config: str = None, is_cloud: bool = False):
         self.base_dir = base_dir
         self.project_id = project_id
         self.project_job_id = project_job_id
@@ -130,6 +130,11 @@ class PipelineExecutor:
         self.current_node = None
         self.job_status = job_status
         self.current_node_from_config = current_node_from_config
+        self.is_cloud = is_cloud
+        self.background_executors = []  # Keep track of background executors
+        self.background_jobs_count = 0  # Track number of active background jobs
+        self.background_jobs_completed = 0  # Track completed background jobs
+        self.background_jobs_lock = threading.Lock()  # Thread-safe counter updates
         
         # For resuming: flag to track if we've reached the resume point
         self.resume_mode = (job_status == "running" and current_node_from_config is not None)
@@ -153,6 +158,18 @@ class PipelineExecutor:
         # Initialize log file
         self.init_log()
 
+    # ✅ Helper method to conditionally notify cloud manager
+    def _notify_cloud(self, completion_type: str):
+        """Triggers scientiflow-cloud commands if in cloud mode."""
+        if not self.is_cloud:
+            return
+        try:
+            cmd = ["scientiflow-cloud", f"--completed-job-{completion_type}", str(self.project_job_id)]
+            subprocess.run(cmd, check=False)
+            logger.info(f"Sent {completion_type} signal to cloud manager.")
+        except Exception as e:
+            print(f"[ERROR] Cloud notification failed: {e}")
+
     def init_log(self):
         """Initialize the log file."""
         try:
@@ -185,6 +202,8 @@ class PipelineExecutor:
             body = {"project_job_id": self.project_job_id, "terminal_output": terminal_output}
             make_auth_request(endpoint="/agent-application/update-terminal-output", method="POST", data=body, error_message="Unable to update terminal output!")
             printer.print_message("[+] Terminal output updated successfully.", style="bold green")
+            # ✅ TRIGGER: FULL COMPLETION
+            self._notify_cloud("fully")
         except Exception as e:
             print(f"[ERROR] Failed to update terminal output: {e}")
 
@@ -241,6 +260,20 @@ class PipelineExecutor:
             raise SystemExit("[ERROR] Pipeline execution terminated due to an unexpected error.")
 
 
+    def wait_for_background_jobs(self):
+        """Wait for all background jobs to complete."""
+        import time
+        if self.background_jobs_count > 0:
+            printer.print_message(f"[INFO] Waiting for {self.background_jobs_count} background job(s) to complete...", style="bold yellow")
+            # ✅ TRIGGER: PARTIAL COMPLETION (Frees GPU/Primary CPU while waiting for cleanup)
+            self._notify_cloud("partially")
+            while True:
+                with self.background_jobs_lock:
+                    if self.background_jobs_completed >= self.background_jobs_count:
+                        break
+                time.sleep(0.5)  # Check every 500ms
+            
+            printer.print_message("[INFO] All background jobs completed.", style="bold green")
 
     def dfs(self, node: str):
         """Perform Depth-First Search (DFS) for executing pipeline nodes."""
@@ -353,7 +386,6 @@ class PipelineExecutor:
         """Start executing the pipeline."""
         # Use job status from configuration instead of API call
         current_status = self.job_status
-        
         if current_status == "running":
             # Job is already running, resume from start but skip until current node
             current_node_id = self.current_node_from_config
@@ -384,7 +416,8 @@ class PipelineExecutor:
         self.update_terminal_output()
 
 # External function to initiate the pipeline execution
-def decode_and_execute_pipeline(base_dir: str, project_id: int, project_job_id: int, project_title: str, job_dir_name: str, nodes: List[Dict[str, Any]], edges: List[Dict[str, str]], environment_variables: Dict[str, str], start_node: str = None, end_node: str = None, job_status: str = None, current_node_from_config: str = None):
+def decode_and_execute_pipeline(base_dir: str, project_id: int, project_job_id: int, project_title: str, job_dir_name: str, nodes: List[Dict[str, Any]], edges: List[Dict[str, str]], environment_variables: Dict[str, str], start_node: str = None, end_node: str = None, job_status: str = None, current_node_from_config: str = None, is_cloud: bool = False):
     """Initialize and execute the pipeline."""
-    executor = PipelineExecutor(base_dir, project_id, project_job_id, project_title, job_dir_name, nodes, edges, environment_variables, start_node, end_node, job_status, current_node_from_config)
-    executor.decode_and_execute_pipeline()
+    executor = PipelineExecutor(base_dir, project_id, project_job_id, project_title, job_dir_name, nodes, edges, environment_variables, start_node, end_node, job_status, current_node_from_config, is_cloud)
+    executor.decode_and_execute_pipeline()
+

{scientiflow_cli-0.4.13 → scientiflow_cli-0.4.16}/scientiflow_cli/services/executor.py

@@ -37,7 +37,7 @@ def execute_jobs(job_ids: list[int] = None, parallel: bool = False, is_cloud: bo
             if matching_jobs:
                 if is_cloud:
                     for job in matching_jobs:
-                        job['project_title'] = job['project']['id'] + '_' + job['project']['project_title']
+                        job['project']['project_title'] = str(job['project']['id']) + '_' + job['project']['project_title']
                         if 'server' not in job or job['server'] is None:
                             job['server'] = {'base_directory':None}
                         job['server']['base_directory'] = get_base_directory()
@@ -172,8 +172,8 @@ def execute_single_job(job: dict, is_cloud: bool = False) -> None:
             printer.print_success(f"[+] Resuming execution for job ID: {project_job_id}")
         else:
             printer.print_success(f"[+] Starting execution for job ID: {project_job_id}")
-            
-        decode_and_execute_pipeline(base_dir, project_id, project_job_id, project_title, job_dir_name, nodes, edges, environment_variables, start_node=start_node, end_node=end_node, job_status=job_status, current_node_from_config=current_node_from_config)
+
+        decode_and_execute_pipeline(base_dir, project_id, project_job_id, project_title, job_dir_name, nodes, edges, environment_variables, start_node=start_node, end_node=end_node, job_status=job_status, current_node_from_config=current_node_from_config, is_cloud=is_cloud)
         printer.print_success(f"[+] Execution completed for job ID: {project_job_id}")
 
     except ValueError as value_err: