Pynndescent knn

CHANGELOG.md

@@ -38,7 +38,11 @@
 
 * `reference/cellranger_mkgtf`: Rename `reference/mkgtf` to `reference/cellranger_mkgtf` (PR #846).
 
-* `interactive/run_cellxgene` and `interactive/run_cirrocumulus` were removed (PR #796).
+* `labels_transfer/xgboost`: Align interface with new annotation workflow
+  - Store label probabilities instead of uncertainties
+  - Take `.h5mu` format as an input instead of `.h5ad`
+
+* `labels_transfer/knn`: delete outdated component due to its functionality now implemented in `labels_transfer/pynndescent_knn`
 
 * `reference/build_cellranger_arc_reference`: a default value of "output" is now specified for the argument `--genome`, inline with `reference/build_cellranger_reference` component. Additionally, providing a value for `--organism` is no longer required and its default value of `Homo Sapiens` has been removed (PR #864).
 
@@ -83,6 +87,8 @@
 
 * `dataflow/split_h5mu` component: Added a component to split a single h5mu file into multiple h5mu files based on the values of an .obs column (PR #824).
 
+* `labels_transfer/pynndescent_knn`: component: Added a component for KNN classification based on a PyNNDescent neighborhood graph (PR #830).
+
 * `workflows/test_workflows/ingestion` components & `workflows/ingestion`: Added standalone components for integration testing of ingestion workflows (PR #801). 
 
 * `workflows/ingestion/make_reference`: Add additional arguments passed through to the STAR and BD Rhapsody reference components (PR #846).

src/labels_transfer/api/common_arguments.yaml

@@ -1,4 +1,3 @@
-
   argument_groups:
     - name: Input dataset (query) arguments
       arguments:
@@ -25,7 +24,7 @@
                     obsm:
                       - type: "double"
                         name: "features"
-                        example: X_integrated_scanvi
+                        example: X_scvi
                         required: false
                         description: |
                           The embedding to use for the classifier's inference. Override using the `--input_obsm_features` argument. If not provided, the `.X` slot will be used instead.
@@ -41,52 +40,59 @@
           description: |
             The `.obsm` key of the embedding to use for the classifier's inference. If not provided, the `.X` slot will be used instead.
             Make sure that embedding was obtained in the same way as the reference embedding (e.g. by the same model or preprocessing).
-          example: X_integrated_scanvi
+          example: X_scvi
+
     - name: Reference dataset arguments
       arguments:
         - name: "--reference"
           type: "file"
           required: false
           description: "The reference data to train classifiers on."
-          example: https://zenodo.org/record/6337966/files/HLCA_emb_and_metadata.h5ad
+          example: reference.h5mu
           info:
             label: "Reference"
             file_format:
-              type: h5ad
-              X:
-                type: double
-                name: features
-                required: false
-                description: |
-                  The expression data to use for the classifier's training, if `--input_obsm_features` argument is not provided.
-              obsm:
-                - type: "double"
-                  name: "features"
-                  example: X_integrated_scanvi
-                  description: |
-                    The embedding to use for the classifier's training. Override using the `--reference_obsm_features` argument.
-                    Make sure that embedding was obtained in the same way as the query embedding (e.g. by the same model or preprocessing).
-                  required: true
-              obs:
-                - type: "string"
-                  name: "targets"
-                  multiple: true
-                  example: [ ann_level_1, ann_level_2, ann_level_3, ann_level_4, ann_level_5, ann_finest_level ]
-                  description: "The target labels to transfer. Override using the `--reference_obs_targets` argument."
+              type: h5mu
+              mod:
+                rna:
+                  description: "Modality in AnnData format containing RNA data."
                   required: true
+                  slots:
+                    X:
+                      type: double
+                      name: features
+                      required: false
+                      description: |
+                        The expression data to use for the classifier's training, if `--input_obsm_features` argument is not provided.
+                    obsm:
+                      - type: "double"
+                        name: "features"
+                        example: X_scvi
+                        description: |
+                          The embedding to use for the classifier's training. Override using the `--reference_obsm_features` argument.
+                          Make sure that embedding was obtained in the same way as the query embedding (e.g. by the same model or preprocessing).
+                        required: true
+                    obs:
+                      - type: "string"
+                        name: "targets"
+                        multiple: true
+                        example: [ ann_level_1, ann_level_2, ann_level_3, ann_level_4, ann_level_5, ann_finest_level ]
+                        description: "The target labels to transfer. Override using the `--reference_obs_targets` argument."
+                        required: true
         - name: "--reference_obsm_features"
           type: string
-          required: true
+          required: false
           description: |
-            The `.obsm` key of the embedding to use for the classifier's training.
+            The `.obsm` key of the embedding to use for the classifier's training. If not provided, the `.X` slot will be used instead.
             Make sure that embedding was obtained in the same way as the query embedding (e.g. by the same model or preprocessing).
-          default: X_integrated_scanvi
+          example: X_scvi
         - name: "--reference_obs_targets"
           type: string
           default: [ ann_level_1, ann_level_2, ann_level_3, ann_level_4, ann_level_5, ann_finest_level ]
           required: false
           multiple: true
-          description: The `.obs` key of the target labels to tranfer.
+          description: The `.obs` key(s) of the target labels to tranfer.
+
     - name: Outputs
       arguments:
         - name: "--output"
@@ -108,22 +114,16 @@
                       description: "The predicted labels. Override using the `--output_obs_predictions` argument."
                       required: true
                     - type: "double"
-                      name: "uncertainty"
-                      description: "The uncertainty of the predicted labels. Override using the `--output_obs_uncertainty` argument."
+                      name: "probability"
+                      description: "The probability of the predicted labels. Override using the `--output_obs_probability` argument."
                       required: false
                   obsm:
                     - type: "double"
-                      name: "X_integrated_scanvi"
+                      name: "X_scvi"
                       description: 
                         The embedding used for the classifier's inference. Could have any name,
                         specified by `input_obsm_features` argument."
                       required: false
-                  uns:
-                    - type: "string"
-                      name: "parameters"
-                      example: labels_tranfer
-                      description: Additional information about the parameters used for the label transfer.
-                      required: true
         - name: "--output_obs_predictions"
           type: string
           example: 
@@ -133,17 +133,19 @@
             In which `.obs` slots to store the predicted information.
             If provided, must have the same length as `--reference_obs_targets`.
             If empty, will default to the `reference_obs_targets` combined with the `"_pred"` suffix.
-        - name: "--output_obs_uncertainty"
+        - name: "--output_obs_probability"
           type: string
           example: 
           required: false
           multiple: true
           description: |
-            In which `.obs` slots to store the uncertainty of the predictions.
+            In which `.obs` slots to store the probability of the predictions.
             If provided, must have the same length as `--reference_obs_targets`.
-            If empty, will default to the `reference_obs_targets` combined with the `"_uncertainty"` suffix.
-        - name: "--output_uns_parameters"
+            If empty, will default to the `reference_obs_targets` combined with the `"_probability"` suffix.
+        - name: "--output_compression"
           type: string
-          default: "labels_transfer"
+          description: |
+            The compression format to be used on the output h5mu object.
+          choices: ["gzip", "lzf"]
           required: false
-          description: "The `.uns` key to store additional information about the parameters used for the label transfer."
+          example: "gzip"