fixs: details are as follows:

1. if matrix input is gem, transfer gem to gef
2. single cell demo refer to StereoSeq_ssDNA_param.json

cellbin2/config/StereoSeq_singlecell_dapi_param.json

@@ -0,0 +1,46 @@
+{
+    "image_process": {
+        "0": {
+            "file_path": "/media/Data/dzh/data/cellbin2/demo_data/Y00788P9/Y00788P9_aligned_image.tif",
+            "tech_type": "ssDNA",
+            "chip_detect": false,
+            "quality_control": false,
+            "registration": {
+                "fixed_image": 1,
+                "trackline": true,
+                "reuse": -1
+            },
+            "tissue_segmentation": false,
+            "cell_segmentation": true,
+            "correct_r": 10,
+            "channel_align": -1
+        },
+        "1": {
+            "file_path": "/media/Data/dzh/data/cellbin2/demo_data/Y00788P9/Y00788P9.gem.gz",
+            "tech_type": "Transcriptomics",
+            "chip_detect": false,
+            "quality_control": false,
+            "registration": {
+                "fixed_image": -1,
+                "trackline": true,
+                "reuse": -1
+            },
+            "tissue_segmentation": true,
+            "cell_segmentation": false,
+            "correct_r": 0,
+            "channel_align": -1
+        }
+    },
+    "molecular_classify": {
+        "0": {
+            "exp_matrix": 1,
+            "cell_mask": [0],
+            "extra_method": ""
+        }
+    },
+    "analysis": {
+        "cluster": true,
+        "annotation": false,
+        "report": true
+    }
+}

cellbin2/matrix/matrix.py

@@ -212,12 +212,29 @@ def adjust_mask_shape(gef_path, mask_path):
     return new_path
 
 
+def gem_to_gef(gem_path, gef_path):
+    from gefpy.bgef_writer_cy import generate_bgef
+    generate_bgef(input_file=gem_path,
+                  bgef_file=gef_path,
+                  stromics="Transcriptomics",
+                  n_thread=8,
+                  bin_sizes=[1],
+                  )
+
+
 def save_cell_bin_data(src_path: str, dst_path: str, cell_mask: str):
     """ 获取：单细胞数据（mask可来自配准图像 or 矩阵自身 ） """
     src_path = str(src_path)
     dst_path = str(dst_path)
     cell_mask = str(cell_mask)
     from gefpy import cgef_writer_cy
+    if src_path.endswith(".gem.gz"):
+        gef_path = os.path.join(os.path.dirname(dst_path), os.path.basename(src_path).replace(".gem.gz", ".raw.gef"))
+        if os.path.exists(gef_path):
+            src_path = gef_path
+        else:
+            gem_to_gef(src_path, gef_path)
+            src_path = gef_path
     if src_path.endswith(".gef"):
         cell_mask = adjust_mask_shape(gef_path=src_path, mask_path=cell_mask)
     cgef_writer_cy.generate_cgef(dst_path, src_path, cell_mask, [256, 256])

cellbin2/modules/extract/image_extract.py

@@ -154,41 +154,47 @@ def extract4transform(
                 trans_chip_box_info.ScaleX, trans_chip_box_info.ScaleY = 1.0, 1.0
                 trans_chip_box_info.Rotation = 0.0
                 self._channel_image.Stitch.TransformChipBBox.update(trans_chip_box_info)
-
-        if self._image_file.tissue_segmentation and not self._naming.transform_tissue_mask_raw.exists():
-            tissue_mask = self._tissue_segmentation(
-                image_path=self._naming.transformed_image,
-                res_path=self._naming.transform_tissue_mask_raw
-            )
-        if self._image_file.cell_segmentation and not self._naming.transform_cell_mask_raw.exists():
-            cell_mask = self._cell_segmentation(
-                image_path=self._naming.transformed_image,
-                res_path=self._naming.transform_cell_mask_raw
-            )
+        final_tissue_mask = None
+        final_cell_mask = None
+        tissue_mask = None
+        cell_mask = None
         if self._image_file.tissue_segmentation:
-            if 'tissue_mask' not in locals():
+            if not self._naming.transform_tissue_mask_raw.exists():
+                tissue_mask = self._tissue_segmentation(
+                    image_path=self._naming.transformed_image,
+                    res_path=self._naming.transform_tissue_mask_raw
+                )
+            else:
                 tissue_mask_p = self._naming.transform_tissue_mask_raw
                 tissue_mask = cbimread(tissue_mask_p, only_np=True)
-            if self._image_file.cell_segmentation:
-                if 'cell_mask' not in locals():
-                    cell_mask_p = self._naming.transform_cell_mask_raw
-                    cell_mask = cbimread(cell_mask_p, only_np=True)
-                input_data = MaskManagerInfo(
-                    tissue_mask=tissue_mask,
-                    cell_mask=cell_mask,
-                    chip_box=self._channel_image.Stitch.TransformChipBBox.get(),
-                    method=1,
-                    stain_type=self._image_file.tech
-                )
-                btcm = BestTissueCellMask.get_best_tissue_cell_mask(input_data=input_data)
-                final_tissue_mask = btcm.best_tissue_mask
-                final_cell_mask = btcm.best_cell_mask
-                cbimwrite(
-                    output_path=self._naming.transform_cell_mask,
-                    files=final_cell_mask
+            final_tissue_mask = tissue_mask
+        if self._image_file.cell_segmentation:
+            if not self._naming.transform_cell_mask_raw.exists():
+                cell_mask = self._cell_segmentation(
+                    image_path=self._naming.transformed_image,
+                    res_path=self._naming.transform_cell_mask_raw
                 )
             else:
-                final_tissue_mask = tissue_mask
+                cell_mask_p = self._naming.transform_cell_mask_raw
+                cell_mask = cbimread(cell_mask_p, only_np=True)
+            final_cell_mask = cell_mask
+        if tissue_mask is not None and cell_mask is not None:
+            input_data = MaskManagerInfo(
+                tissue_mask=tissue_mask,
+                cell_mask=cell_mask,
+                chip_box=self._channel_image.Stitch.TransformChipBBox.get(),
+                method=1,
+                stain_type=self._image_file.tech
+            )
+            btcm = BestTissueCellMask.get_best_tissue_cell_mask(input_data=input_data)
+            final_tissue_mask = btcm.best_tissue_mask
+            final_cell_mask = btcm.best_cell_mask
+        if final_cell_mask is not None:
+            cbimwrite(
+                output_path=self._naming.transform_cell_mask,
+                files=final_cell_mask
+            )
+        if final_tissue_mask is not None:
             cbimwrite(
                 output_path=self._naming.transform_tissue_mask,
                 files=final_tissue_mask

cellbin2/modules/extract/matrix_extract.py

@@ -41,20 +41,24 @@ def extract4matrix(self, m_naming: naming.DumpPipelineFileNaming):
         cell_mask_path = m_naming.final_nuclear_mask
         tissue_mask_path = m_naming.final_tissue_mask
         cell_correct_mask_path = m_naming.final_cell_mask
+        c_inp = None
         if Path(tissue_mask_path).exists():
             save_tissue_bin_data(
                 self._image_file.file_path,
                 str(self._naming.tissue_bin_matrix),
                 tissue_mask_path,
             )
+            c_inp = self._naming.tissue_bin_matrix
+        if c_inp is None:
+            c_inp = self._image_file.file_path
         if Path(cell_mask_path).exists():
             save_cell_bin_data(
-                self._naming.tissue_bin_matrix,
+                c_inp,
                 str(self._naming.cell_bin_matrix),
                 cell_mask_path)
         if Path(cell_correct_mask_path).exists():
             save_cell_bin_data(
-                self._naming.tissue_bin_matrix,
+                c_inp,
                 str(self._naming.cell_correct_bin_matrix),
                 cell_correct_mask_path
             )