Fix test data

sapporo-wes · Sep 20, 2022 · dd97550 · dd97550
1 parent 37af986
commit dd97550
Show file tree

Hide file tree

Showing 15 changed files with 783 additions and 35 deletions.
diff --git a/README.md b/README.md
@@ -113,7 +113,7 @@ outputs/                    service_info.json           workflow_engine_params.t
 run.pid                     start_time.txt              yevis-metadata.yml
 
 # Execute sapporo/ro_crate.py script
-$ docker run --rm -v $PWD:$PWD -w $PWD ghcr.io/sapporo-wes/sapporo-service:latest python3 /app/sapporo/ro_crate.py $PWD
+$ docker run --rm -v /var/run/docker.sock:/var/run/docker.sock -v $PWD:$PWD -w $PWD ghcr.io/sapporo-wes/sapporo-service:latest python3 /app/sapporo/ro_crate.py $PWD
 ```
 
 ## Development

diff --git a/tests/example_crate/gatk_1st.json b/tests/example_crate/gatk_1st.json
@@ -84,7 +84,7 @@
             "author": {
                 "@id": "https://orcid.org/0000-0003-2765-0049"
             },
-            "datePublished": "2022-09-17T15:58:25+00:00",
+            "datePublished": "2022-09-20T14:09:26+00:00",
             "description": "RO-Crate of Sapporo's run execution results.",
             "hasPart": [
                 {
@@ -2529,7 +2529,7 @@
             ],
             "cpuArchitecture": "x86_64",
             "cpuCount": 4,
-            "freeDiskSpace": 138291744768,
+            "freeDiskSpace": 138369884160,
             "gid": 0,
             "inDocker": true,
             "os": "Linux",
@@ -4162,7 +4162,7 @@
                 }
             ],
             "pid": 12,
-            "runId": "work",
+            "runId": "gatk_1st",
             "startTime": "2022-09-08T09:30:03",
             "state": "COMPLETE",
             "stderr": {
@@ -4265,6 +4265,28 @@
             "uid": 1000,
             "url": "http://localhost:1122/runs/c71fb9c2-b6e8-48f4-9350-c9b6ab4647dc/data/outputs/G97753.NA12878.vcf.idx"
         },
+        {
+            "@id": "#samtools",
+            "@type": "SoftwareApplication",
+            "name": "samtools",
+            "version": "1.15.1--h1170115_0"
+        },
+        {
+            "@id": "#8a8cb096-c6ac-46ea-8da1-8818e562748c",
+            "@type": [
+                "FileStats"
+            ],
+            "duplicateRate": 0.3063459248661424,
+            "duplicateReads": 719878,
+            "generatedBy": {
+                "@id": "#samtools"
+            },
+            "mappedRate": 1.0,
+            "mappedReads": 2349886,
+            "totalReads": 2349886,
+            "unmappedRate": 0.0,
+            "unmappedReads": 0
+        },
         {
             "@id": "outputs/G97753.NA12878.bam",
             "@type": [
@@ -4281,6 +4303,9 @@
             "gid": 1000,
             "mode": "-rw-r--r--",
             "sha512": "5bb2600f09cdbb7847e2bac06ebf1323f97a9243a0323c5429c375ed262965334952f3e4eda904096aa7ea6f1f79a4c05f5443785257ee981bbaff65ce2d531c",
+            "stats": {
+                "@id": "#8a8cb096-c6ac-46ea-8da1-8818e562748c"
+            },
             "uid": 1000,
             "url": "http://localhost:1122/runs/c71fb9c2-b6e8-48f4-9350-c9b6ab4647dc/data/outputs/G97753.NA12878.bam"
         },
@@ -4331,6 +4356,24 @@
             "uid": 1000,
             "url": "http://localhost:1122/runs/c71fb9c2-b6e8-48f4-9350-c9b6ab4647dc/data/outputs/G97753.NA12878.vcf"
         },
+        {
+            "@id": "#vcftools",
+            "@type": "SoftwareApplication",
+            "name": "vcftools",
+            "version": "0.1.16--pl5321h9a82719_6"
+        },
+        {
+            "@id": "#c80f6b40-0f8e-4a32-96f4-8e4b3678f8e1",
+            "@type": [
+                "FileStats"
+            ],
+            "generatedBy": {
+                "@id": "#vcftools"
+            },
+            "indelsCount": 2,
+            "snpsCount": 16,
+            "variantCount": 18
+        },
         {
             "@id": "outputs/splitAndPassOnly.vcf",
             "@type": [
@@ -4348,6 +4391,9 @@
             "lineCount": 96,
             "mode": "-rw-r--r--",
             "sha512": "9278bb53e11183d0be5fcb45c5eea49235bbbc02da9d2181c4c32bd13550a8059f0348e3db4314de3778214dd584271063c97d60b89687328255a5c786ac3fa4",
+            "stats": {
+                "@id": "#c80f6b40-0f8e-4a32-96f4-8e4b3678f8e1"
+            },
             "uid": 1000,
             "url": "http://localhost:1122/runs/c71fb9c2-b6e8-48f4-9350-c9b6ab4647dc/data/outputs/splitAndPassOnly.vcf"
         },
@@ -4368,6 +4414,22 @@
             "uid": 1000,
             "url": "http://localhost:1122/runs/c71fb9c2-b6e8-48f4-9350-c9b6ab4647dc/data/outputs/per_base_coverage.tsv"
         },
+        {
+            "@id": "#fe047f14-c581-485e-823f-1150f5b9d4f4",
+            "@type": [
+                "FileStats"
+            ],
+            "duplicateRate": 0.3063243148927087,
+            "duplicateReads": 723304,
+            "generatedBy": {
+                "@id": "#samtools"
+            },
+            "mappedRate": 0.9999944944088605,
+            "mappedReads": 2361223,
+            "totalReads": 2361236,
+            "unmappedRate": 5.505591139555724e-06,
+            "unmappedReads": 13
+        },
         {
             "@id": "outputs/G97753.NA12878.realigned.bam",
             "@type": [
@@ -4384,6 +4446,9 @@
             "gid": 1000,
             "mode": "-rw-r--r--",
             "sha512": "057c8d8eeb508e0792e39fd45fbfab4eecb9c15b601a4d9fdc540ae21e55aaaddcb8079978fa97efe799db0fddfa6b792cac9db662bb0d121da403acceda8781",
+            "stats": {
+                "@id": "#fe047f14-c581-485e-823f-1150f5b9d4f4"
+            },
             "uid": 1000,
             "url": "http://localhost:1122/runs/c71fb9c2-b6e8-48f4-9350-c9b6ab4647dc/data/outputs/G97753.NA12878.realigned.bam"
         },

diff --git a/tests/example_crate/gatk_2nd.json b/tests/example_crate/gatk_2nd.json
@@ -84,7 +84,7 @@
             "author": {
                 "@id": "https://orcid.org/0000-0003-2765-0049"
             },
-            "datePublished": "2022-09-17T15:59:23+00:00",
+            "datePublished": "2022-09-20T14:10:50+00:00",
             "description": "RO-Crate of Sapporo's run execution results.",
             "hasPart": [
                 {
@@ -2529,7 +2529,7 @@
             ],
             "cpuArchitecture": "x86_64",
             "cpuCount": 4,
-            "freeDiskSpace": 138291736576,
+            "freeDiskSpace": 138361393152,
             "gid": 0,
             "inDocker": true,
             "os": "Linux",
@@ -4162,7 +4162,7 @@
                 }
             ],
             "pid": 12,
-            "runId": "work",
+            "runId": "gatk_2nd",
             "startTime": "2022-09-08T08:59:34",
             "state": "COMPLETE",
             "stderr": {
@@ -4265,6 +4265,28 @@
             "uid": 1000,
             "url": "http://localhost:1122/runs/cc91b422-25a2-4721-8d66-477ef4e141a7/data/outputs/G97753.NA12878.vcf.idx"
         },
+        {
+            "@id": "#samtools",
+            "@type": "SoftwareApplication",
+            "name": "samtools",
+            "version": "1.15.1--h1170115_0"
+        },
+        {
+            "@id": "#74251a63-fbc5-4694-97d3-bed9a43c6a30",
+            "@type": [
+                "FileStats"
+            ],
+            "duplicateRate": 0.3063459248661424,
+            "duplicateReads": 719878,
+            "generatedBy": {
+                "@id": "#samtools"
+            },
+            "mappedRate": 1.0,
+            "mappedReads": 2349886,
+            "totalReads": 2349886,
+            "unmappedRate": 0.0,
+            "unmappedReads": 0
+        },
         {
             "@id": "outputs/G97753.NA12878.bam",
             "@type": [
@@ -4281,6 +4303,9 @@
             "gid": 1000,
             "mode": "-rw-r--r--",
             "sha512": "2ca1e340a4eec22d5d254d5e5c30bd7fea5215fd09f4b424553933b88a61f971b59049d550054fa911c4cc686b08e82a7106bab44a4ac7cadc8e4a525718be43",
+            "stats": {
+                "@id": "#74251a63-fbc5-4694-97d3-bed9a43c6a30"
+            },
             "uid": 1000,
             "url": "http://localhost:1122/runs/cc91b422-25a2-4721-8d66-477ef4e141a7/data/outputs/G97753.NA12878.bam"
         },
@@ -4331,6 +4356,24 @@
             "uid": 1000,
             "url": "http://localhost:1122/runs/cc91b422-25a2-4721-8d66-477ef4e141a7/data/outputs/G97753.NA12878.vcf"
         },
+        {
+            "@id": "#vcftools",
+            "@type": "SoftwareApplication",
+            "name": "vcftools",
+            "version": "0.1.16--pl5321h9a82719_6"
+        },
+        {
+            "@id": "#c2c735ed-405b-4ac6-aba1-eeaf31e2d3a6",
+            "@type": [
+                "FileStats"
+            ],
+            "generatedBy": {
+                "@id": "#vcftools"
+            },
+            "indelsCount": 2,
+            "snpsCount": 16,
+            "variantCount": 18
+        },
         {
             "@id": "outputs/splitAndPassOnly.vcf",
             "@type": [
@@ -4348,6 +4391,9 @@
             "lineCount": 96,
             "mode": "-rw-r--r--",
             "sha512": "e869e0002f11260c7e2573f11d84a67f281301ccc20744d84f11409552214154dddfad3acf09aacf2e9af3d985a41147ac64563fda64750767576dffaf343d67",
+            "stats": {
+                "@id": "#c2c735ed-405b-4ac6-aba1-eeaf31e2d3a6"
+            },
             "uid": 1000,
             "url": "http://localhost:1122/runs/cc91b422-25a2-4721-8d66-477ef4e141a7/data/outputs/splitAndPassOnly.vcf"
         },
@@ -4368,6 +4414,22 @@
             "uid": 1000,
             "url": "http://localhost:1122/runs/cc91b422-25a2-4721-8d66-477ef4e141a7/data/outputs/per_base_coverage.tsv"
         },
+        {
+            "@id": "#80c75c41-a993-4f09-a6cc-e11faa5f5581",
+            "@type": [
+                "FileStats"
+            ],
+            "duplicateRate": 0.3063243148927087,
+            "duplicateReads": 723304,
+            "generatedBy": {
+                "@id": "#samtools"
+            },
+            "mappedRate": 0.9999944944088605,
+            "mappedReads": 2361223,
+            "totalReads": 2361236,
+            "unmappedRate": 5.505591139555724e-06,
+            "unmappedReads": 13
+        },
         {
             "@id": "outputs/G97753.NA12878.realigned.bam",
             "@type": [
@@ -4384,6 +4446,9 @@
             "gid": 1000,
             "mode": "-rw-r--r--",
             "sha512": "2b818b96876ccceae64a715f4ce08a0254265598c98f20130dec8e3fed7b241f8cb4ef3bed3da611f59460b55744ebb6b137c4e31806e48f3232a105eb180ab7",
+            "stats": {
+                "@id": "#80c75c41-a993-4f09-a6cc-e11faa5f5581"
+            },
             "uid": 1000,
             "url": "http://localhost:1122/runs/cc91b422-25a2-4721-8d66-477ef4e141a7/data/outputs/G97753.NA12878.realigned.bam"
         },

diff --git a/tests/example_crate/gatk_mac.json b/tests/example_crate/gatk_mac.json
@@ -84,7 +84,7 @@
             "author": {
                 "@id": "https://orcid.org/0000-0003-2765-0049"
             },
-            "datePublished": "2022-09-17T16:01:36+00:00",
+            "datePublished": "2022-09-20T14:13:10+00:00",
             "description": "RO-Crate of Sapporo's run execution results.",
             "hasPart": [
                 {
@@ -637,7 +637,7 @@
             ],
             "cpuArchitecture": "x86_64",
             "cpuCount": 4,
-            "freeDiskSpace": 138291724288,
+            "freeDiskSpace": 138361294848,
             "gid": 0,
             "inDocker": true,
             "os": "Linux",
@@ -1323,7 +1323,7 @@
             ],
             "name": "Sapporo test result",
             "pid": 615,
-            "runId": "work",
+            "runId": "gatk_mac",
             "startTime": "2022-09-14T05:42:22",
             "state": "EXECUTOR_ERROR",
             "stderr": {