merge develop

.github/workflows/website.yml

@@ -8,7 +8,7 @@ on:
       # 'main' triggers updates to 'sleap.ai', all others to 'sleap.ai/develop'
       - main
       - develop
-      - liezl/update-intallation-docs-1.4.1 # again!
+      - liezl/add-channels-for-pip-conda-env
     paths:
       - "docs/**"
       - "README.rst"

docs/guides/bonsai.md

@@ -0,0 +1,75 @@
+(bonsai)=
+
+# Using Bonsai with SLEAP
+
+Bonsai is a visual language for reactive programming and currently supports SLEAP models.
+
+:::{note}
+Currently Bonsai supports only single instance, top-down and top-down-id SLEAP models.
+:::
+
+### Exporting a SLEAP trained model
+
+Before we can import a trained model into Bonsai, we need to use the {code}`sleap-export` command to convert the model to a format supported by Bonsai. For example, to export a top-down-id model, the command is as follows:
+
+```bash
+sleap-export -m centroid/model/folder/path -m top_down_id/model/folder/path -e exported/model/path
+```
+
+Please refer to the {ref}`sleap-export` docs for more details on using the command.
+
+This will generate the necessary `.pb` file and other information files required by Bonsai. In this example, these files were saved to the specified `exported/model/path` folder.
+
+The `exported/model/path` folder will have a structure like the following:
+
+```plaintext
+exported/model/path
+├── centroid_config.json
+├── confmap_config.json
+├── frozen_graph.pb
+└── info.json
+```
+
+### Installing Bonsai and necessary packages
+
+1. Install Bonsai. See the [Bonsai installation instructions](https://bonsai-rx.org/docs/articles/installation.html). 
+
+2. Download and add the necessary packages for Bonsai to run with SLEAP. See the official [Bonsai SLEAP documentation](https://github.com/bonsai-rx/sleap?tab=readme-ov-file#bonsai---sleap) for more information.
+
+### Using Bonsai SLEAP modules
+
+Once you have Bonsai installed with the required packages, you should be able to open the Bonsai application. The workflow must have a source module `FileCapture` which can be found in the toolbox search in the workflow editor. Provide the path to the video that was used to train the SLEAP model in the `FileName` field of the module.
+
+![Bonsai FileCapture module](../_static/bonsai-filecapture.jpg)
+
+#### Top-down model
+The top-down model requires both the `PredictCentroids` and the `PredictPoses` modules.
+
+The `PredictCentroids` module will predict the centroids of detections. There are two fields inside the `PredictCentroids` module: the `ModelFileName` field and the `TrainingConfig` field. The `TrainingConfig` field expects the path to the training config JSON file for the centroid model. The `ModelFileName` field expects the path to the `frozen_graph.pb` file in the `exported/model/path` folder.
+
+![Bonsai PredictCentroids module](../_static/bonsai-predictcentroids.jpg)
+
+The `PredictPoses` module will predict the instances of detections. Similar to the `PredictCentroid` module, there are two fields inside the `PredictPoses` module: the `ModelFileName` field and the `TrainingConfig` field. The `TrainingConfig` field expects the path to the training config JSON file for the centered instance model. The `ModelFileName` field expects the path to the `frozen_graph.pb` file in the `exported/model/path` folder.
+
+![Bonsai PredictPoses module](../_static/bonsai-predictposes.jpg)
+
+#### Top-Down-ID model
+The `PredictPoseIdentities` module will predict the instances with identities. This module has two fields: the `ModelFileName` field and the `TrainingConfig` field. The `TrainingConfig` field expects the path to the training config JSON file for the top-down-id model. The `ModelFileName` field expects the path to the `frozen_graph.pb` file in the `exported/model/path` folder.
+
+![Bonsai PredictPoseIdentities module](../_static/bonsai-predictposeidentities.jpg)
+
+#### Single instance model
+The `PredictSinglePose` module will predict the poses for single instance models. This module also has two fields: the `ModelFileName` field and the `TrainingConfig` field. The `TrainingConfig` field expects the path to the training config JSON file for the single instance model. The `ModelFileName` field expects the path to the `frozen_graph.pb` file in the `exported/model/path` folder.
+
+### Connecting the modules
+Right-click on the `FileCapture` module and select **Create Connection**. Now click on the required SLEAP module to complete the connection. 
+
+![Bonsai module connection ](../_static/bonsai-connection.jpg)
+
+Once it is done, the workflow in Bonsai will look something like the following:
+
+![Bonsai.SLEAP workflow](../_static/bonsai-workflow.jpg)
+
+Now you can click the green start button to run the workflow and you can add more modules to analyze and visualize the results in Bonsai.
+
+For more documentation on various modules and workflows, please refer to the [official Bonsai docs](https://bonsai-rx.org/docs/articles/editor.html).

docs/guides/index.md

@@ -30,6 +30,10 @@
 
 {ref}`remote-inference` when you trained models and you want to run inference on a different machine using a **command-line interface**.
 
+## SLEAP with Bonsai
+
+{ref}`bonsai` when you want to analyze the trained SLEAP model to visualize the poses, centroids and identities for further visual analysis.
+
 ```{toctree}
 :hidden: true
 :maxdepth: 2
@@ -44,4 +48,5 @@ proofreading
 colab
 custom-training
 remote
+bonsai
 ```

docs/installation.md

@@ -27,7 +27,7 @@ local:
    Installation requires entering commands in a terminal. To open one:
    ````{tabs}
       ```{tab} Windows
-         Open the *Start menu* and search for the *Anaconda Prompt* (if using Miniconda) or the *Command Prompt* if not. 
+         Open the *Start menu* and search for the *Anaconda Prompt* (if using Miniconda) or the *Command Prompt* if not.
          ```{note}
          On Windows, our personal preference is to use alternative terminal apps like [Cmder](https://cmder.net) or [Windows Terminal](https://aka.ms/terminal).
          ```
@@ -66,7 +66,6 @@ If you don't have a `conda` package manager installation, here are some quick in
 
 Miniforge is a minimal installer for conda that includes the `conda` package manager and is maintained by the [conda-forge](https://conda-forge.org) community. The only difference between Miniforge and Miniconda is that Miniforge uses the `conda-forge` channel by default, which provides a much wider selection of community-maintained packages.
 
-
 ````{tabs}
    ```{group-tab} Windows
       Open a new PowerShell terminal (does not need to be admin) and enter:
@@ -135,20 +134,20 @@ This is a minimal installer for conda that includes the `conda` package manager
 
 See the [Miniconda website](https://docs.anaconda.com/free/miniconda/) for up-to-date installation instructions if the above instructions don't work for your system.
 
-
 (installation-methods)=
+
 ## Installation methods
 
 SLEAP can be installed three different ways: via {ref}`conda package<condapackage>`, {ref}`conda from source<condasource>`, or {ref}`pip package<pippackage>`. Select one of the methods below to install SLEAP. We recommend {ref}`conda package<condapackage>`.
 
-````{tabs}
+`````{tabs}
    ```{tab} conda package
       **This is the recommended installation method**.
       ````{tabs}
          ```{group-tab} Windows and Linux
             ```bash
             conda create -y -n sleap -c conda-forge -c nvidia -c sleap -c anaconda sleap=1.4.1a2
-            ```  
+            ```
             ```{note}
             - This comes with CUDA to enable GPU support. All you need is to have an NVIDIA GPU and [updated drivers](https://nvidia.com/drivers).
             - If you already have CUDA installed on your system, this will not conflict with it.
@@ -222,7 +221,7 @@ SLEAP can be installed three different ways: via {ref}`conda package<condapackag
                ````{tabs}
                   ```{group-tab} NVIDIA GPU
                      ```bash
-                     conda create --name sleap pip python=3.7.12 cudatoolkit=11.3 cudnn=8.2
+                     conda create --name sleap pip python=3.7.12 cudatoolkit=11.3 cudnn=8.2 -c conda-forge -c nvidia
                      ```
                   ```
                   ```{group-tab} CPU or other GPU
@@ -256,7 +255,7 @@ SLEAP can be installed three different ways: via {ref}`conda package<condapackag
          ```
          ````
    ```
-````
+`````
 
 ## Testing that things are working
 

sleap/config/frame_range_form.yaml

@@ -0,0 +1,13 @@
+main:
+
+  - name: min_frame_idx
+    label: Minimum frame index
+    type: int
+    range: 1,1000000
+    default: 1
+
+  - name: max_frame_idx
+    label: Maximum frame index
+    type: int
+    range: 1,1000000
+    default: 1000

sleap/gui/app.py

@@ -656,17 +656,18 @@ def prev_vid():
             key="edge style",
         )
 
+        # XXX
         add_submenu_choices(
             menu=viewMenu,
             title="Node Marker Size",
-            options=(1, 2, 4, 6, 8, 12),
+            options=prefs["node marker sizes"],
             key="marker size",
         )
 
         add_submenu_choices(
             menu=viewMenu,
             title="Node Label Size",
-            options=(6, 12, 18, 24, 36),
+            options=prefs["node label sizes"],
             key="node label size",
         )
 
@@ -804,6 +805,12 @@ def new_instance_menu_action():
             "Delete Predictions beyond Max Instances...",
             self.commands.deleteInstanceLimitPredictions,
         )
+        add_menu_item(
+            labelMenu,
+            "delete frame limit predictions",
+            "Delete Predictions beyond Frame Limit...",
+            self.commands.deleteFrameLimitPredictions,
+        )
 
         ### Tracks Menu ###
 
@@ -873,6 +880,8 @@ def new_instance_menu_action():
             "Point Displacement (max)",
             "Primary Point Displacement (sum)",
             "Primary Point Displacement (max)",
+            "Tracking Score (mean)",
+            "Tracking Score (min)",
             "Instance Score (sum)",
             "Instance Score (min)",
             "Point Score (sum)",
@@ -1331,7 +1340,7 @@ def updateStatusMessage(self, message: Optional[str] = None):
                 message += f" [Hidden] Press '{hide_key}' to toggle."
                 self.statusBar().setStyleSheet("color: red")
             else:
-                self.statusBar().setStyleSheet("color: black")
+                self.statusBar().setStyleSheet("")
 
         self.statusBar().showMessage(message)
 
@@ -1406,6 +1415,8 @@ def _set_seekbar_header(self, graph_name: str):
             "Point Displacement (max)": data_obj.get_point_displacement_series,
             "Primary Point Displacement (sum)": data_obj.get_primary_point_displacement_series,
             "Primary Point Displacement (max)": data_obj.get_primary_point_displacement_series,
+            "Tracking Score (mean)": data_obj.get_tracking_score_series,
+            "Tracking Score (min)": data_obj.get_tracking_score_series,
             "Instance Score (sum)": data_obj.get_instance_score_series,
             "Instance Score (min)": data_obj.get_instance_score_series,
             "Point Score (sum)": data_obj.get_point_score_series,
@@ -1419,7 +1430,7 @@ def _set_seekbar_header(self, graph_name: str):
         else:
             if graph_name in header_functions:
                 kwargs = dict(video=self.state["video"])
-                reduction_name = re.search("\\((sum|max|min)\\)", graph_name)
+                reduction_name = re.search("\\((sum|max|min|mean)\\)", graph_name)
                 if reduction_name is not None:
                     kwargs["reduction"] = reduction_name.group(1)
                 series = header_functions[graph_name](**kwargs)

sleap/gui/commands.py

@@ -49,6 +49,7 @@ class which inherits from `AppCommand` (or a more specialized class such as
 from sleap.gui.dialogs.merge import MergeDialog, ReplaceSkeletonTableDialog
 from sleap.gui.dialogs.message import MessageDialog
 from sleap.gui.dialogs.missingfiles import MissingFilesDialog
+from sleap.gui.dialogs.frame_range import FrameRangeDialog
 from sleap.gui.state import GuiState
 from sleap.gui.suggestions import VideoFrameSuggestions
 from sleap.instance import Instance, LabeledFrame, Point, PredictedInstance, Track
@@ -494,6 +495,10 @@ def deleteInstanceLimitPredictions(self):
         """Gui for deleting instances beyond some number in each frame."""
         self.execute(DeleteInstanceLimitPredictions)
 
+    def deleteFrameLimitPredictions(self):
+        """Gui for deleting instances beyond some frame number."""
+        self.execute(DeleteFrameLimitPredictions)
+
     def completeInstanceNodes(self, instance: Instance):
         """Adds missing nodes to given instance."""
         self.execute(AddMissingInstanceNodes, instance=instance)
@@ -2472,6 +2477,36 @@ def ask(cls, context: CommandContext, params: dict) -> bool:
             return super().ask(context, params)
 
 
+class DeleteFrameLimitPredictions(InstanceDeleteCommand):
+    @staticmethod
+    def get_frame_instance_list(context: CommandContext, params: Dict):
+        """Called from the parent `InstanceDeleteCommand.ask` method.
+
+        Returns:
+            List of instances to be deleted.
+        """
+        instances = []
+        # Select the instances to be deleted
+        for lf in context.labels.labeled_frames:
+            if lf.frame_idx < (params["min_frame_idx"] - 1) or lf.frame_idx > (
+                params["max_frame_idx"] - 1
+            ):
+                instances.extend([(lf, inst) for inst in lf.instances])
+        return instances
+
+    @classmethod
+    def ask(cls, context: CommandContext, params: Dict) -> bool:
+        current_video = context.state["video"]
+        dialog = FrameRangeDialog(
+            title="Delete Instances in Frame Range...", max_frame_idx=len(current_video)
+        )
+        results = dialog.get_results()
+        if results:
+            params["min_frame_idx"] = results["min_frame_idx"]
+            params["max_frame_idx"] = results["max_frame_idx"]
+            return super().ask(context, params)
+
+
 class TransposeInstances(EditCommand):
     topics = [UpdateTopic.project_instances, UpdateTopic.tracks]
 

sleap/gui/dataviews.py

@@ -15,20 +15,17 @@
 
 """
 
-from qtpy import QtCore, QtWidgets, QtGui
-
-import numpy as np
 import os
-
 from operator import itemgetter
+from pathlib import Path
+from typing import Any, Callable, List, Optional
 
-from typing import Any, Callable, Dict, List, Optional, Type
+import numpy as np
+from qtpy import QtCore, QtGui, QtWidgets
 
-from sleap.gui.state import GuiState
 from sleap.gui.commands import CommandContext
-from sleap.gui.color import ColorManager
-from sleap.io.dataset import Labels
-from sleap.instance import LabeledFrame, Instance
+from sleap.gui.state import GuiState
+from sleap.instance import LabeledFrame
 from sleap.skeleton import Skeleton
 
 
@@ -386,10 +383,25 @@ def getSelectedRowItem(self) -> Any:
 
 
 class VideosTableModel(GenericTableModel):
-    properties = ("filename", "frames", "height", "width", "channels")
-
-    def item_to_data(self, obj, item):
-        return {key: getattr(item, key) for key in self.properties}
+    properties = (
+        "name",
+        "filepath",
+        "frames",
+        "height",
+        "width",
+        "channels",
+    )
+
+    def item_to_data(self, obj, item: "Video"):
+        data = {}
+        for property in self.properties:
+            if property == "name":
+                data[property] = Path(item.filename).name
+            elif property == "filepath":
+                data[property] = str(Path(item.filename).parent)
+            else:
+                data[property] = getattr(item, property)
+        return data
 
 
 class SkeletonNodesTableModel(GenericTableModel):