Adding COSMX data export pipeline

.github/workflows/python-app.yml

@@ -5,7 +5,7 @@ name: Python application
 
 on:
   push:
-    branches: [ master, fix_dag25]
+    branches: [ master, cosmx_may25]
   pull_request:
     branches: [ "master" ]
 

data/validation_schema/cosmx_protein_count_file_validation_schema.json

@@ -0,0 +1,49 @@
+{
+    "$schema": "https://json-schema.org/draft/2020-12/json-schema-core.html#",
+    "id": "https://github.com/imperial-genomics-facility/data-management-python/tree/master/data/validation_schema/cosmx_protein_count_file_validation_schema.json#",
+    "title": "IGF COSMX PROTEIN count file validation schema",
+    "description": "Schema for validation of COSMX PROTEIN count file",
+    "type" : "array",
+    "version": "1.0.0",
+    "minItems": 1,
+    "uniqueItems": true,
+    "items": {
+        "type": "object",
+        "properties" : {
+            "fov_id" : {
+              "type" : "integer"
+            },
+            "mean_fluorescence_intensity" : {
+              "type" : "integer"
+            },
+            "mean_unique_genes_per_cell" : {
+              "type" : "number"
+            },
+            "number_non_empty_cells" : {
+              "type" : "integer"
+            },
+            "pct_non_empty_cells" : {
+              "type" : "number"
+            },
+            "percentile_10_fluorescence_intensity" : {
+              "type" : "number"
+            },
+            "percentile_90_fluorescence_intensity" : {
+              "type" : "number"
+            },
+            "fluorescence_intensity_mean_igg_control_intensity": {
+              "type" : "number"
+            }
+        },
+        "required": [
+          "fov_id",
+          "mean_fluorescence_intensity",
+          "mean_unique_genes_per_cell",
+          "number_non_empty_cells",
+          "pct_non_empty_cells",
+          "percentile_10_fluorescence_intensity",
+          "percentile_90_fluorescence_intensity",
+          "fluorescence_intensity_mean_igg_control_intensity"
+        ]
+    }
+}

data/validation_schema/cosmx_rna_count_file_validation_schema.json

@@ -0,0 +1,49 @@
+{
+    "$schema": "https://json-schema.org/draft/2020-12/json-schema-core.html#",
+    "id": "https://github.com/imperial-genomics-facility/data-management-python/tree/master/data/validation_schema/cosmx_rna_count_file_validation_schema.json#",
+    "title": "IGF COSMX RNA count file validation schema",
+    "description": "Schema for validation of COSMX RNA count file",
+    "type" : "array",
+    "version": "1.0.0",
+    "minItems": 1,
+    "uniqueItems": true,
+    "items": {
+        "type": "object",
+        "properties" : {
+            "fov_id" : {
+              "type" : "integer"
+            },
+            "mean_transcript_per_cell" : {
+              "type" : "number"
+            },
+            "mean_unique_genes_per_cell" : {
+              "type" : "number"
+            },
+            "number_non_empty_cells" : {
+              "type" : "integer"
+            },
+            "pct_non_empty_cells" : {
+              "type" : "number"
+            },
+            "percentile_90_transcript_per_cell" : {
+              "type" : "number"
+            },
+            "percentile_10_transcript_per_cell" : {
+              "type" : "number"
+            },
+            "mean_negprobe_counts_per_cell": {
+              "type" : "number"
+            }
+        },
+        "required": [
+          "fov_id",
+          "mean_transcript_per_cell",
+          "mean_unique_genes_per_cell",
+          "number_non_empty_cells",
+          "pct_non_empty_cells",
+          "percentile_90_transcript_per_cell",
+          "percentile_10_transcript_per_cell",
+          "mean_negprobe_counts_per_cell"
+        ]
+    }
+}

data/validation_schema/minimal_metadata_validation.json

@@ -0,0 +1,55 @@
+{
+    "$schema": "http://json-schema.org/draft-06/schema#",
+    "id": "https://github.com/imperial-genomics-facility/data-management-python/tree/master/data/validation_schema/milimal_metadata_validation.json#",
+    "title": "IGF minimal metadata validation schema",
+    "description": "Schema for validation of minimal metadata input file",
+    "type" : "array",
+    "version": "1.0.0",
+    "minItems": 1,
+    "uniqueItems": true,
+    "items": {
+        "type": "object",
+        "properties" : {
+            "project_igf_id" : {
+              "type" : "string",
+              "pattern": "^IGF[a-zA-Z0-9-_]+$",
+              "maxLength": 50
+            },
+            "deliverable" : {
+              "type" : "string",
+              "enum" : ["FASTQ",
+                        "COSMX"]
+            },
+            "name" : {
+              "type" : "string",
+              "pattern": "^\\w+\\s\\w+\\s?\\w+?$",
+              "maxLength": 30
+            },
+            "username" : {
+              "type" : "string",
+              "pattern": "^\\S+$|^$",
+              "maxLength": 20
+            },
+            "email_id" : {
+              "type" : "string",
+              "pattern": "\\S+\\@\\S+\\.\\w+",
+              "maxLength": 40
+            },
+            "sample_igf_id" : {
+              "type" : "string",
+              "pattern": "^IGF[a-zA-Z0-9-_]+$",
+              "maxLength": 20
+            },
+            "species_name" : {
+              "type" : "string",
+              "maxLength": 50
+            }
+        },
+        "required": [
+          "project_igf_id",
+          "deliverable",
+          "name",
+          "email_id",
+          "username"]
+    }
+}