fix: srp index error

.github/workflows/build_push.yml

@@ -0,0 +1,43 @@
+on:
+    workflow_call:
+      inputs:
+        PUSH:
+          required: true
+          type: boolean
+          default: false
+
+jobs:
+    build_push:
+      runs-on: ubuntu-latest
+      permissions:
+        packages: write
+        contents: read
+        id-token: write
+      steps:
+        - uses: actions/checkout@v4
+        - name: Set up QEMU
+          uses: docker/setup-qemu-action@v3
+        - name: Set up Docker Buildx
+          uses: docker/setup-buildx-action@v3
+        - name: Login to GitHub Container Registry
+          uses: docker/login-action@v3
+          with:
+            registry: ghcr.io
+            username: ${{ github.actor }}
+            password: ${{ github.token }}
+        - if: "${{ inputs.PUSH == false }}"
+          name: Build only
+          uses: docker/build-push-action@v6
+          with:
+            context: .
+            platforms: linux/amd64
+            push: ${{ inputs.PUSH }}
+            tags: ghcr.io/${{ github.repository }}:latest
+        - if: "${{ inputs.PUSH == true }}"
+          name: Build and push
+          uses: docker/build-push-action@v6
+          with:
+            context: .
+            platforms: linux/amd64,linux/arm64
+            push: ${{ inputs.PUSH }}
+            tags: ghcr.io/${{ github.repository }}:${{ github.ref_name }},ghcr.io/${{ github.repository }}:latest

.github/workflows/on_pr.yml

@@ -0,0 +1,11 @@
+name: CI
+on: pull_request
+jobs:
+  ci:
+    permissions:
+      packages: write
+      contents: read
+      id-token: write
+    uses: ./.github/workflows/build_push.yml
+    with:
+      PUSH: false

.github/workflows/on_push_tag.yml

@@ -0,0 +1,14 @@
+name: CD
+on:
+  push:
+    tags:
+    - '*'
+jobs:
+  cd:
+    permissions:
+      packages: write
+      contents: read
+      id-token: write
+    uses: ./.github/workflows/build_push.yml
+    with:
+      PUSH: true

modules/sparse_ref_panel.py

@@ -34,6 +34,7 @@
 To select a part of the chromosome by position, use range:
 sparse_matrix = ref_panel.range(start_pos, end_pos)
 """
+
 import os
 import subprocess
 from io import BytesIO
@@ -88,6 +89,10 @@ def __getitem__(self, key: Tuple[Union[int, list, slice]]) -> sparse.csc_matrix:
             raise TypeError("Both variant and haplotype slices must be provided")
         # handle single row
         if isinstance(key[0], int):
+            if key[0] > self.n_variants - 1:
+                raise IndexError(
+                    f"Index {key[0]} out of range for {self.n_variants} variants"
+                )
             return self._load_haplotypes(key[0] // self.chunk_size)[
                 key[0] % self.chunk_size, key[1]
             ]
@@ -102,7 +107,9 @@ def __getitem__(self, key: Tuple[Union[int, list, slice]]) -> sparse.csc_matrix:
                     ]
                 )
             row_stop = (
-                min([key[0].stop, self.n_variants]) if key[0].stop is not None else self.n_variants
+                min([key[0].stop, self.n_variants - 1])
+                if key[0].stop is not None
+                else self.n_variants - 1
             )
             chunks = list(
                 range(
@@ -144,7 +151,9 @@ def __getitem__(self, key: Tuple[Union[int, list, slice]]) -> sparse.csc_matrix:
 
         chunks, splits = np.unique(rows // self.chunk_size, return_index=True)
         if any(chunk not in self.chunks[:, 0] for chunk in chunks):
-            raise IndexError("Variants index out of range")
+            raise IndexError(
+                f"Index {key[0]} out of range for {self.n_variants} variants"
+            )
 
         chunk_idx = np.split(rows % self.chunk_size, splits[1:])
 

selphi.py

@@ -62,6 +62,15 @@ def selphi(
     ref_panel = SparseReferencePanel(
         str(ref_base_path.with_suffix(".srp")), cache_size=2
     )
+    # Confirm that reference panel files match
+    n_pbwt_samples = ref_base_path.with_suffix(".samples").read_text().count("\n")
+    n_pbwt_vars = ref_base_path.with_suffix(".sites").read_text().count("\n")
+    if n_pbwt_samples != ref_panel.n_samples or n_pbwt_vars != ref_panel.n_variants:
+        raise ValueError(
+            f"pbwt reference panel has {n_pbwt_samples} samples and {n_pbwt_vars} "
+            f"variants while sparse reference panel has {ref_panel.n_samples} samples "
+            f"and {ref_panel.n_variants} variants"
+        )
 
     # Load target samples and variants
     vcf_obj = cyvcf2.VCF(targets_path)