Merge pull request #2 from UW-GAC/feature/add-genotype-inventory

Genotype inventory

.dockstore.yml

@@ -5,3 +5,13 @@ workflows:
      primaryDescriptorPath: /primed_phenotype_inventory.wdl
      testParameterFiles:
          - /primed_phenotype_inventory.json
+   - name: primed_genotype_inventory
+     subclass: WDL
+     primaryDescriptorPath: /primed_genotype_inventory.wdl
+     testParameterFiles:
+         - /primed_genotype_inventory.json
+   - name: primed_inventories
+     subclass: WDL
+     primaryDescriptorPath: /primed_inventories.wdl
+     testParameterFiles:
+         - /primed_inventories.json

primed_genotype_inventory.json

@@ -0,0 +1,10 @@
+{
+    "primed_genotype_inventory.workspaces": {
+        "primed-data-prevent-1/PRIMED_ARIC_DBGAP_PHS000280_V8_P2_HMB-IRB": "ARIC",
+        "primed-data-dprism-1/PRIMED_RPGEH_DBGAP_PHS000788_V2_P3_HMB-IRB-NPU": "GERA, RPGEH",
+        "primed-data-topmed-1/PRIMED_CARDIA_TOPMED_DBGAP_PHS001612_V1_P1_HMB-IRB": "CARDIA"
+    },
+    "primed_genotype_inventory.output_workspace_namespace": "primed-adrienne",
+    "primed_genotype_inventory.output_workspace_name": "primed-sandbox",
+    "primed_genotype_inventory.output_table": "genotype_inventory"
+}

primed_genotype_inventory.wdl

@@ -0,0 +1,44 @@
+version 1.0
+
+workflow primed_genotype_inventory {
+    input {
+        Map[String, String] workspaces
+        String output_workspace_name
+        String output_workspace_namespace
+        String output_table
+    }
+
+    call write_primed_genotype_inventory_table {
+        input: workspaces = workspaces,
+                output_workspace_name = output_workspace_name,
+                output_workspace_namespace = output_workspace_namespace,
+                output_table = output_table
+    }
+
+    meta {
+          author: "Adrienne stilp"
+          email: "[email protected]"
+    }
+}
+
+task write_primed_genotype_inventory_table {
+    input {
+        Map[String, String] workspaces
+        String output_workspace_name
+        String output_workspace_namespace
+        String output_table
+    }
+
+    command <<<
+        set -e
+        Rscript /usr/local/primed-inventory-workflows/write_primed_genotype_inventory_table.R \
+            --workspaces-file ~{write_map(workspaces)} \
+            --output-workspace-name ~{output_workspace_name} \
+            --output-workspace-namespace ~{output_workspace_namespace} \
+            --output-table-name ~{output_table}
+    >>>
+
+    runtime {
+        docker: "uwgac/primed-inventory-workflows:0.2.0"
+    }
+}

primed_inventories.json

@@ -0,0 +1,9 @@
+{
+    "primed_inventories.workspaces": {
+        "primed-data-prevent-1/PRIMED_ARIC_DBGAP_PHS000280_V8_P2_HMB-IRB": "ARIC",
+        "primed-data-dprism-1/PRIMED_RPGEH_DBGAP_PHS000788_V2_P3_HMB-IRB-NPU": "GERA, RPGEH",
+        "primed-data-topmed-1/PRIMED_CARDIA_TOPMED_DBGAP_PHS001612_V1_P1_HMB-IRB": "CARDIA"
+    },
+    "primed_inventories.output_workspace_namespace": "primed-adrienne",
+    "primed_inventories.output_workspace_name": "primed-sandbox"
+}

primed_inventories.wdl

@@ -0,0 +1,31 @@
+version 1.0
+
+import "primed_phenotype_inventory.wdl" as phenotype_inventory
+import "primed_genotype_inventory.wdl" as genotype_inventory
+
+workflow primed_inventories {
+    input {
+        Map[String, String] workspaces
+        String output_workspace_name
+        String output_workspace_namespace
+    }
+
+    call phenotype_inventory.write_primed_phenotype_inventory_table {
+        input: workspaces = workspaces,
+                output_workspace_name = output_workspace_name,
+                output_workspace_namespace = output_workspace_namespace,
+                output_table = "phenotype_inventory"
+    }
+
+    call genotype_inventory.write_primed_genotype_inventory_table {
+        input: workspaces = workspaces,
+                output_workspace_name = output_workspace_name,
+                output_workspace_namespace = output_workspace_namespace,
+                output_table = "genotype_inventory"
+    }
+
+    meta {
+          author: "Adrienne stilp"
+          email: "[email protected]"
+    }
+}

write_primed_genotype_inventory_table.R

@@ -0,0 +1,115 @@
+library(argparser)
+library(AnVIL)
+library(AnvilDataModels)
+library(dplyr)
+library(readr)
+library(tidyr)
+library(stringr)
+
+
+argp <- arg_parser("write_genotype_inventory_table.R", description="Write genotype inventory table for shared workspaces.")
+argp <- add_argument(argp, "--workspaces-file", help="2-column tsv file with (workspace, studies)")
+argp <- add_argument(argp, "--output-workspace-namespace", help="Namespace of the AnVIL workspace to write the genotype inventory table to.")
+argp <- add_argument(argp, "--output-workspace-name", help="Name of the AnVIL workspace to write the genotype inventory table to.")
+argp <- add_argument(argp, "--output-table-name", help="Name of the data table to write the genotype inventory table to.")
+argv <- parse_args(argp)
+
+# Read in the workspaces.
+x <- read_tsv(argv$workspaces_file, col_names=c("workspace", "studies"))
+# x <- tribble(
+#   ~workspace, ~studies,
+#   "primed-data-prevent-1/PRIMED_ARIC_DBGAP_PHS000280_V8_P2_HMB-IRB", "ARIC",
+#   "primed-data-dprism-1/PRIMED_RPGEH_DBGAP_PHS000788_V2_P3_HMB-IRB-NPU", "GERA, RPGEH",
+#   "primed-data-topmed-1/PRIMED_CARDIA_TOPMED_DBGAP_PHS001612_V1_P1_HMB-IRB", "CARDIA"
+# )
+
+genotype_tables <- c(
+  "array_dataset",
+  "imputation_dataset",
+  "sequencing_dataset",
+  "simulation_dataset"
+)
+
+# Split workspace into namespace and name.
+workspaces <- x %>%
+  separate(
+    workspace,
+    into=c("workspace_namespace", "workspace_name"),
+    sep="/",
+    remove=FALSE
+  )
+
+# Just a check:
+print(workspaces)
+
+# Loop over workspaces and pull the genotype inventory information.
+results_list <- list()
+for (i in seq_along(workspaces$workspace)) {
+  print(paste("Processing workspace:", workspaces$workspace[i]))
+  # Loop over the different genotype tables.
+  workspace_results_list <- list()
+  for (input_table_name in genotype_tables) {
+
+    workspace = workspaces$workspace[i]
+    workspace_namespace = workspaces$workspace_namespace[i]
+    workspace_name = workspaces$workspace_name[i]
+
+    tables <- avtables(namespace=workspace_namespace, name=workspace_name)
+    if (input_table_name %in% tables$table) {
+      x <- avtable(input_table_name, namespace=workspace_namespace, name=workspace_name)
+      # Subset to and rename the id column.
+      id_column_name = quo_name(paste0(input_table_name, "_id"))
+      x <- x %>%
+        select(
+          dataset_id = !!id_column_name,
+          reference_assembly,
+          sample_set_id
+        )
+      # Pull the sample set table and calculate the number of samples
+      number_of_samples <- avtable("sample_set", namespace=workspace_namespace, name=workspace_name) %>%
+        unnest_set_table() %>%
+        count(sample_set_id, name="n_samples")
+      x <- x %>% left_join(number_of_samples, by="sample_set_id")
+    }
+    else {
+      x = tibble()
+    }
+    workspace_results_list[[input_table_name]] <- x
+  }
+  results_list[[workspace]] <- bind_rows(workspace_results_list, .id="genotype_dataset_table")
+}
+
+# Combine the results into a single data frame.
+results <- bind_rows(results_list, .id="workspace") %>%
+  left_join(workspaces, by="workspace")
+
+# Set up output workspace info.
+# output_workspace = avworkspace() # This will be different when we actually run the script.
+# output_workspace_namespace = str_split_1(output_workspace, pattern="/")[1]
+# output_workspace_name = str_split_1(output_workspace, pattern="/")[2]
+# output_table_name <- "tmp_genotype_inventory"
+output_workspace_namespace = argv$output_workspace_namespace
+output_workspace_name = argv$output_workspace_name
+output_table_name = argv$output_table_name
+
+id_column_name = quo_name(paste0(output_table_name, "_id"))
+results <- results %>%
+  select(dataset_id, everything()) %>%
+  rename(!!id_column_name := dataset_id) %>%
+  # We separated workspace into namespace and name, so we don't need it anymore.
+  select(-workspace)
+print(results)
+
+# Delete the table before writing the new data, if it already exists.
+tables <- avtables(namespace=output_workspace_namespace, name=output_workspace_name)
+if (output_table_name %in% tables$table) {
+  original_results <- avtable(output_table_name, namespace=output_workspace_namespace, name=output_workspace_name)
+  avtable_delete_values(output_table_name, original_results[[id_column_name]], namespace=output_workspace_namespace, name=output_workspace_name)
+}
+
+tables <- list(tmp=results) %>%
+  setNames(output_table_name)
+
+# Write the new results the table.
+# Note: anvil_import_tables will check job status and timeout after an hour (by default).
+anvil_import_tables(tables, namespace=output_workspace_namespace, name=output_workspace_name, overwrite=TRUE)