PyPI - workbench - Versions diffs - 0.8.218__tar.gz → 0.8.224__tar.gz - Mend

workbench 0.8.218tar.gz → 0.8.224tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (839) hide show

{workbench-0.8.218 → workbench-0.8.224}/PKG-INFO RENAMED Viewed

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: workbench
-Version: 0.8.218
+Version: 0.8.224
 Summary: Workbench: A Dashboard and Python API for creating and deploying AWS SageMaker Model Pipelines
 Author-email: SuperCowPowers LLC <support@supercowpowers.com>
 License: MIT License
@@ -54,6 +54,7 @@ Requires-Dist: requests>=2.26.0
 Requires-Dist: rdkit>=2024.9.5
 Requires-Dist: mordredcommunity>=2.0.6
 Requires-Dist: workbench-bridges>=0.1.16
+Requires-Dist: cleanlab>=2.9.0
 Provides-Extra: ui
 Requires-Dist: plotly>=6.0.0; extra == "ui"
 Requires-Dist: dash>=3.0.0; extra == "ui"
@@ -71,18 +72,8 @@ Requires-Dist: flake8; extra == "dev"
 Requires-Dist: black; extra == "dev"
 Provides-Extra: all
 Requires-Dist: networkx>=3.2; extra == "all"
-Requires-Dist: plotly>=6.0.0; extra == "all"
-Requires-Dist: dash>=3.0.0; extra == "all"
-Requires-Dist: dash-bootstrap-components>=1.6.0; extra == "all"
-Requires-Dist: dash-bootstrap-templates>=1.3.0; extra == "all"
-Requires-Dist: dash_ag_grid; extra == "all"
-Requires-Dist: tabulate>=0.9.0; extra == "all"
-Requires-Dist: pytest; extra == "all"
-Requires-Dist: pytest-sugar; extra == "all"
-Requires-Dist: coverage; extra == "all"
-Requires-Dist: pytest-cov; extra == "all"
-Requires-Dist: flake8; extra == "all"
-Requires-Dist: black; extra == "all"
+Requires-Dist: workbench[ui]; extra == "all"
+Requires-Dist: workbench[dev]; extra == "all"
 Dynamic: license-file

{workbench-0.8.218 → workbench-0.8.224}/applications/aws_dashboard/app.py RENAMED Viewed

@@ -1,12 +1,13 @@
 """Workbench Dashboard: A Workbench Web Application for viewing and managing Workbench Artifacts"""
-from dash import Dash, html, dcc, page_container
+from dash import Dash, html, dcc, page_container, Input, Output, State, ALL
 import dash_bootstrap_components as dbc
 from dash import page_registry
 # Workbench Imports
 from workbench.utils.plugin_manager import PluginManager
 from workbench.utils.theme_manager import ThemeManager
+from workbench.web_interface.components.settings_menu import SettingsMenu
 # Set up the logging
 import logging
@@ -37,9 +38,43 @@ app = Dash(
 # Register the CSS route in the ThemeManager
 tm.register_css_route(app)
+# Custom index string to sync localStorage theme to cookie on page load
+app.index_string = """
+<!DOCTYPE html>
+<html>
+    <head>
+        {%metas%}
+        <title>{%title%}</title>
+        {%favicon%}
+        {%css%}
+        <script>
+            // Sync localStorage theme to cookie on page load (before Flask renders)
+            (function() {
+                var theme = localStorage.getItem('wb_theme');
+                if (theme) {
+                    document.cookie = 'wb_theme=' + theme + '; path=/; max-age=31536000';
+                }
+            })();
+        </script>
+    </head>
+    <body>
+        {%app_entry%}
+        <footer>
+            {%config%}
+            {%scripts%}
+            {%renderer%}
+        </footer>
+    </body>
+</html>
+"""
 # Note: The 'server' object is required for running the app with NGINX/uWSGI
 server = app.server
+# Create the settings menu component
+settings_menu = SettingsMenu()
+settings_menu_id = "settings-menu"
 # For Multi-Page Applications, we need to create a 'page container' to hold all the pages
 plugin_info_id = "plugin-pages-info"
 app.layout = html.Div(
@@ -47,11 +82,32 @@ app.layout = html.Div(
         # URL for subpage navigation (jumping to feature_sets, models, etc.)
         dcc.Location(id="url", refresh="callback-nav"),
         dcc.Store(id=plugin_info_id, data={}),
+        # Settings menu in top-right corner
+        html.Div(
+            settings_menu.create_component(settings_menu_id),
+            style={"position": "absolute", "top": "25px", "right": "20px", "zIndex": 1000},
+        ),
         dbc.Container([page_container], fluid=True, className="dbc dbc-ag-grid"),
     ],
     **{"data-bs-theme": tm.data_bs_theme()},
 )
+# Clientside callback for theme switching (stores in localStorage, sets cookie, reloads page)
+app.clientside_callback(
+    settings_menu.get_clientside_callback_code(settings_menu_id),
+    Output(f"{settings_menu_id}-dummy", "data"),
+    Input({"type": f"{settings_menu_id}-theme-item", "theme": ALL}, "n_clicks"),
+    State({"type": f"{settings_menu_id}-theme-item", "theme": ALL}, "id"),
+)
+# Clientside callback to update checkmarks based on localStorage
+app.clientside_callback(
+    settings_menu.get_checkmark_callback_code(),
+    Output({"type": f"{settings_menu_id}-checkmark", "theme": ALL}, "children"),
+    Input(f"{settings_menu_id}-init", "data"),
+    State({"type": f"{settings_menu_id}-checkmark", "theme": ALL}, "id"),
+)
 # Spin up the Plugin Manager
 pm = PluginManager()

workbench-0.8.224/applications/aws_dashboard/pages/models/callbacks.py ADDED Viewed

@@ -0,0 +1,155 @@
+"""Callbacks for the Model Subpage Web User Interface"""
+import logging
+from dash import callback, no_update, Input, Output, State
+from dash.exceptions import PreventUpdate
+from urllib.parse import urlparse, parse_qs
+# Workbench Imports
+from workbench.api import ModelType
+from workbench.web_interface.page_views.models_page_view import ModelsPageView
+from workbench.web_interface.components.plugins.ag_table import AGTable
+from workbench.web_interface.components.plugins.scatter_plot import ScatterPlot
+from workbench.web_interface.components.plugins.confusion_matrix import ConfusionMatrix
+from workbench.cached.cached_model import CachedModel
+# Get the Workbench logger
+log = logging.getLogger("workbench")
+def on_page_load():
+    @callback(
+        Output("models_table", "selectedRows"),
+        Output("models_page_loaded", "data"),
+        Input("url", "href"),
+        Input("models_table", "rowData"),
+        State("models_page_loaded", "data"),
+        prevent_initial_call=True,
+    )
+    def _on_page_load(href, row_data, page_already_loaded):
+        if page_already_loaded:
+            raise PreventUpdate
+        if not href or not row_data:
+            raise PreventUpdate
+        parsed = urlparse(href)
+        if parsed.path != "/models":
+            raise PreventUpdate
+        selected_name = parse_qs(parsed.query).get("name", [None])[0]
+        if not selected_name:
+            return [row_data[0]], True
+        for row in row_data:
+            if row.get("name") == selected_name:
+                return [row], True
+        raise PreventUpdate
+def model_table_refresh(page_view: ModelsPageView, table: AGTable):
+    @callback(
+        [Output(component_id, prop) for component_id, prop in table.properties],
+        Input("models_refresh", "n_intervals"),
+    )
+    def _model_table_refresh(_n):
+        """Return the table data for the Models Table"""
+        page_view.refresh()
+        models = page_view.models()
+        models["name"] = models["Model Group"]
+        models["id"] = range(len(models))
+        return table.update_properties(models)
+def setup_plugin_callbacks(plugins, scatter_plot_plugin: ScatterPlot, confusion_matrix_plugin: ConfusionMatrix):
+    """Set up callbacks for all plugins including model-type-aware visualization switching.
+    Args:
+        plugins: List of all plugins (model_details, shap_plot, scatter_plot, confusion_matrix)
+        scatter_plot_plugin: The ScatterPlot plugin instance (for regression models)
+        confusion_matrix_plugin: The ConfusionMatrix plugin instance (for classification models)
+    """
+    # Register internal callbacks for all plugins
+    for plugin in plugins:
+        plugin.register_internal_callbacks()
+    # Separate plugins by type - ScatterPlot and ConfusionMatrix need special handling
+    standard_plugins = [p for p in plugins if p not in [scatter_plot_plugin, confusion_matrix_plugin]]
+    # Callback to update standard plugins (model_details, shap_plot)
+    @callback(
+        [Output(component_id, prop) for p in standard_plugins for component_id, prop in p.properties],
+        State("model_details-dropdown", "value"),
+        Input("models_table", "selectedRows"),
+    )
+    def update_standard_plugin_properties(inference_run, selected_rows):
+        """Update properties for standard plugins that take a Model object."""
+        if not selected_rows or selected_rows[0] is None:
+            raise PreventUpdate
+        model_name = selected_rows[0]["name"]
+        model = CachedModel(model_name)
+        all_props = []
+        for p in standard_plugins:
+            all_props.extend(p.update_properties(model, inference_run=inference_run))
+        return all_props
+    # Build outputs list for the model plot callback
+    model_plot_outputs = [
+        Output("scatter-plot-container", "style"),
+        Output("confusion-matrix-container", "style"),
+    ]
+    model_plot_outputs.extend([Output(cid, prop) for cid, prop in scatter_plot_plugin.properties])
+    model_plot_outputs.extend([Output(cid, prop) for cid, prop in confusion_matrix_plugin.properties])
+    # Callback to toggle visibility and update ScatterPlot/ConfusionMatrix based on model type
+    @callback(
+        model_plot_outputs,
+        Input("model_details-dropdown", "value"),
+        Input("models_table", "selectedRows"),
+    )
+    def update_model_plot_plugins(inference_run, selected_rows):
+        """Update ScatterPlot or ConfusionMatrix based on model type, toggling visibility."""
+        if not selected_rows or selected_rows[0] is None:
+            raise PreventUpdate
+        model_name = selected_rows[0]["name"]
+        model = CachedModel(model_name)
+        # Determine visibility based on model type
+        is_classifier = model.model_type == ModelType.CLASSIFIER
+        scatter_style = {"display": "none"} if is_classifier else {"display": "block"}
+        confusion_style = {"display": "block"} if is_classifier else {"display": "none"}
+        if is_classifier:
+            # Update ConfusionMatrix, return no_update for ScatterPlot
+            cm_props = confusion_matrix_plugin.update_properties(model, inference_run=inference_run)
+            scatter_props = [no_update] * len(scatter_plot_plugin.properties)
+        else:
+            # Update ScatterPlot with regression data
+            df = model.get_inference_predictions(inference_run)
+            if df is None:
+                raise PreventUpdate
+            # Get target column and compute error for coloring
+            target = model.target()
+            if isinstance(target, list):
+                target = next((t for t in target if t in inference_run), target[0])
+            df["error"] = abs(df["prediction"] - df[target])
+            scatter_props = scatter_plot_plugin.update_properties(
+                df,
+                x=target,
+                y="prediction",
+                color="error",
+                regression_line=True,
+            )
+            cm_props = [no_update] * len(confusion_matrix_plugin.properties)
+        return [scatter_style, confusion_style] + scatter_props + cm_props

{workbench-0.8.218 → workbench-0.8.224}/applications/aws_dashboard/pages/models/layout.py RENAMED Viewed

@@ -18,8 +18,8 @@ project_name = tm.branding().get("project_name", "Workbench")
 def models_layout(
     models_table: AGTable,
     model_details: html.Div,
-    model_plot: dcc.Graph,
-    shap_plot: dcc.Graph,
+    model_plot: html.Div,
+    shap_plot: html.Div,
 ) -> html.Div:
     layout = html.Div(
         children=[
@@ -37,7 +37,7 @@ def models_layout(
                 [
                     # Column 1: Model Details
                     dbc.Col(
-                        dcc.Loading(model_details, type="dot", color="#33aa33", delay_show=300),
+                        dcc.Loading(model_details, type="dot", color="#33aa33", delay_show=1000, delay_hide=100),
                         width=5,
                         className="text-break workbench-container",
                         style={"margin": "20px 0px 0px 0px", "padding": "20px"},

{workbench-0.8.218 → workbench-0.8.224}/applications/aws_dashboard/pages/models/page.py RENAMED Viewed

@@ -1,14 +1,19 @@
 """Models:  A Workbench Web Interface to view, and interact with Models"""
-from dash import register_page
+from dash import register_page, html
 # Local Imports
 from .layout import models_layout
 from . import callbacks
 # Workbench Imports
-from workbench.web_interface.components import model_plot
-from workbench.web_interface.components.plugins import model_details, ag_table, shap_summary_plot
+from workbench.web_interface.components.plugins import (
+    model_details,
+    ag_table,
+    shap_summary_plot,
+    scatter_plot,
+    confusion_matrix,
+)
 from workbench.web_interface.page_views.models_page_view import ModelsPageView
 # Register this page with Dash
@@ -26,8 +31,22 @@ models_table_component = models_table.create_component("models_table", header_co
 my_model_details = model_details.ModelDetails()
 model_details_component = my_model_details.create_component("model_details")
-# Create a Model Plot component to display the model metrics
-model_plot_component = model_plot.ModelPlot().create_component("model_plot")
+# Create ScatterPlot plugin for regression models
+my_scatter_plot = scatter_plot.ScatterPlot()
+scatter_plot_component = my_scatter_plot.create_component("model_scatter_plot")
+# Create ConfusionMatrix plugin for classification models
+my_confusion_matrix = confusion_matrix.ConfusionMatrix()
+confusion_matrix_component = my_confusion_matrix.create_component("model_confusion_matrix")
+# Wrap both in visibility-controlled containers
+model_plot_container = html.Div(
+    [
+        html.Div(scatter_plot_component, id="scatter-plot-container", style={"display": "none"}),
+        html.Div(confusion_matrix_component, id="confusion-matrix-container", style={"display": "none"}),
+    ],
+    id="model-plot-container",
+)
 # Shap summary plot component
 my_shap_plot = shap_summary_plot.ShapSummaryPlot()
@@ -38,7 +57,7 @@ shap_plot_component = my_shap_plot.create_component("shap_plot")
 components = {
     "models_table": models_table_component,
     "model_details": model_details_component,
-    "model_plot": model_plot_component,
+    "model_plot": model_plot_container,
     "shap_plot": shap_plot_component,
 }
@@ -54,12 +73,8 @@ callbacks.on_page_load()
 # Setup our callbacks/connections
 callbacks.model_table_refresh(model_view, models_table)
-# Callback for the model table
-callbacks.update_model_plot_component()
-# Our model details and shap plot are plugins, so we need to set them up
-plugins = [my_model_details, my_shap_plot]
+# All model visualization plugins
+plugins = [my_model_details, my_shap_plot, my_scatter_plot, my_confusion_matrix]
-# Set up callbacks for all the plugins
-if plugins:
-    callbacks.setup_plugin_callbacks(plugins)
+# Set up callbacks for all the plugins (includes scatter plot and confusion matrix)
+callbacks.setup_plugin_callbacks(plugins, my_scatter_plot, my_confusion_matrix)

{workbench-0.8.218 → workbench-0.8.224}/aws_setup/workbench_compute/app.py RENAMED Viewed

@@ -35,6 +35,9 @@ lambda_role_arn = Fn.import_value("WorkbenchCore-LambdaRoleArn")
 existing_vpc_id = cm.get_config("WORKBENCH_VPC_ID")
 subnet_ids = cm.get_config("WORKBENCH_SUBNET_IDS") or []
+# Environment name for notifications (e.g., sandbox, dev, stage, prod)
+environment_name = cm.get_config("WORKBENCH_ENVIRONMENT") or "unknown"
 # Log the configuration for transparency
 print("Configuration:")
 print(f"  WORKBENCH_BUCKET: {workbench_bucket}")
@@ -42,6 +45,7 @@ print(f"  WORKBENCH_BATCH_ROLE_ARN: {batch_role_arn}")
 print(f"  WORKBENCH_LAMBDA_ROLE_ARN: {lambda_role_arn}")
 print(f"  WORKBENCH_VPC_ID: {existing_vpc_id}")
 print(f"  WORKBENCH_SUBNET_IDS: {subnet_ids}")
+print(f"  WORKBENCH_ENVIRONMENT: {environment_name}")
 # Our CDK App and Environment
 app = cdk.App()
@@ -56,6 +60,7 @@ compute_stack = WorkbenchComputeStack(
         workbench_bucket=workbench_bucket,
         batch_role_arn=batch_role_arn,
         lambda_role_arn=lambda_role_arn,
+        environment_name=environment_name,
         existing_vpc_id=existing_vpc_id,
         subnet_ids=subnet_ids,
     ),

workbench-0.8.224/aws_setup/workbench_compute/workbench_compute/lambdas/batch_failure/handler.py ADDED Viewed

@@ -0,0 +1,154 @@
+"""
+Lambda handler for Batch job failure notifications.
+Triggered by EventBridge when a Batch job fails. Fetches CloudWatch logs
+and sends a detailed SNS notification including environment info and error details.
+"""
+import os
+import json
+import boto3
+sns = boto3.client("sns")
+logs = boto3.client("logs")
+def get_log_events(log_stream_name: str, limit: int = 50) -> list[str]:
+    """
+    Fetch recent log events from CloudWatch Logs.
+    Args:
+        log_stream_name: The log stream name (from Batch job details)
+        limit: Maximum number of log events to fetch
+    Returns:
+        List of log message strings
+    """
+    # Batch jobs log to /aws/batch/job
+    log_group = "/aws/batch/job"
+    try:
+        response = logs.get_log_events(
+            logGroupName=log_group,
+            logStreamName=log_stream_name,
+            limit=limit,
+            startFromHead=False,  # Get most recent events
+        )
+        events = response.get("events", [])
+        return [event["message"] for event in events]
+    except logs.exceptions.ResourceNotFoundException:
+        return ["(Log stream not found)"]
+    except Exception as e:
+        return [f"(Error fetching logs: {e})"]
+def format_log_messages(messages: list[str], max_chars: int = 3000) -> str:
+    """
+    Format log messages for SNS, truncating if necessary.
+    Args:
+        messages: List of log message strings
+        max_chars: Maximum total characters to include
+    Returns:
+        Formatted log output string
+    """
+    if not messages:
+        return "(No log messages available)"
+    output = "\n".join(messages)
+    if len(output) > max_chars:
+        output = output[-max_chars:]
+        output = "... (truncated)\n" + output
+    return output
+def lambda_handler(event, context):
+    """
+    Handle Batch job failure event from EventBridge.
+    Extracts job details, fetches CloudWatch logs, and sends SNS notification.
+    """
+    print(f"Received event: {json.dumps(event, indent=2)}")
+    # Extract job details from the event
+    detail = event.get("detail", {})
+    job_name = detail.get("jobName", "Unknown")
+    job_id = detail.get("jobId", "Unknown")
+    status = detail.get("status", "Unknown")
+    status_reason = detail.get("statusReason", "Unknown")
+    # Get container details for log stream
+    container = detail.get("container", {})
+    log_stream_name = container.get("logStreamName", "")
+    # Get environment variables from the job
+    env_vars = {env["name"]: env["value"] for env in container.get("environment", [])}
+    script_path = env_vars.get("ML_PIPELINE_S3_PATH", "N/A")
+    # Environment info
+    environment = os.environ.get("WORKBENCH_ENVIRONMENT", "unknown")
+    region = event.get("region") or os.environ.get("AWS_REGION")
+    topic_arn = os.environ.get("BATCH_FAILURE_TOPIC_ARN")
+    # Fetch CloudWatch logs
+    log_messages = []
+    if log_stream_name:
+        log_messages = get_log_events(log_stream_name)
+    # Build the console URL
+    console_url = f"https://{region}.console.aws.amazon.com/batch/home?region={region}" f"#jobs/fargate/detail/{job_id}"
+    # Build the notification
+    subject = f"Batch Job Failed: {environment}: {job_name}"
+    # Truncate subject if too long (SNS limit is 100 chars)
+    if len(subject) > 100:
+        subject = subject[:97] + "..."
+    message_body = f"""Batch Job Failed!
+Environment: {environment}
+Job Name: {job_name}
+Job ID: {job_id}
+Status: {status}
+Reason: {status_reason}
+Script: {script_path}
+View job details:
+{console_url}
+=== Recent Log Output ===
+{format_log_messages(log_messages)}
+"""
+    # Publish to SNS
+    if topic_arn:
+        try:
+            response = sns.publish(
+                TopicArn=topic_arn,
+                Subject=subject,
+                Message=message_body,
+            )
+            print(f"Published notification to SNS (MessageId: {response['MessageId']})")
+        except Exception as e:
+            print(f"Error publishing to SNS: {e}")
+    else:
+        print("Warning: BATCH_FAILURE_TOPIC_ARN not set, skipping SNS notification")
+        print(f"Would have sent:\nSubject: {subject}\n\n{message_body}")
+    return {
+        "statusCode": 200,
+        "body": json.dumps(
+            {
+                "message": "Batch failure notification sent",
+                "job_name": job_name,
+                "job_id": job_id,
+                "environment": environment,
+            }
+        ),
+    }

workbench 0.8.218__tar.gz → 0.8.224__tar.gz

workbench 0.8.218tar.gz → 0.8.224tar.gz