Support for factors in input data

assets/differentialabundance_report.qmd

@@ -467,10 +467,51 @@ differential_file_suffix <- params$meta$params$differential_file_suffix
 if (is.null(differential_file_suffix)) {
     differential_file_suffix <- paste0(".", params$meta$params$differential_method, ".results.tsv")
 }
-differential_files <- lapply(contrasts$id, function(d) {
-    file.path(params$input_dir, paste0(gsub(" |;", "_", d), differential_file_suffix))
-})
+resolve_differential_file <- function(contrast_id, contrast_index) {
+    candidate_id <- as.character(contrast_id)
+
+    if (!is.na(candidate_id) && nzchar(candidate_id)) {
+        candidate_file <- file.path(params$input_dir, paste0(gsub(" |;", "_", candidate_id), differential_file_suffix))
+        if (file.exists(candidate_file)) {
+            return(candidate_file)
+        }
+    }
+
+    available_files <- list.files(
+        params$input_dir,
+        pattern = paste0(gsub("\\.", "\\\\.", differential_file_suffix), "$"),
+        full.names = TRUE
+    )
+
+    if (length(available_files) == nrow(contrasts)) {
+        return(sort(available_files)[contrast_index])
+    }
+
+    stop(
+        "Could not resolve differential file for contrast row ",
+        contrast_index,
+        " with id '",
+        candidate_id,
+        "'. Available differential files: ",
+        paste(basename(available_files), collapse = ", ")
+    )
+}
+
+differential_files <- mapply(
+    resolve_differential_file,
+    contrasts$id,
+    seq_len(nrow(contrasts)),
+    SIMPLIFY = FALSE
+)
 differential_names <- paste0(contrasts$id)
+missing_differential_names <- is.na(differential_names) | !nzchar(differential_names)
+if (any(missing_differential_names)) {
+    differential_names[missing_differential_names] <- sub(
+        paste0(gsub("\\.", "\\\\.", differential_file_suffix), "$"),
+        "",
+        basename(unlist(differential_files)[missing_differential_names])
+    )
+}
 
 # Initialize vector to store warning messages before merging tables
 warnings_list <- c()

assets/schema_contrasts.json

@@ -78,6 +78,33 @@
                         "additionalProperties": false
                     },
                     "minItems": 1
+                },
+                "variables": {
+                    "type": "object",
+                    "additionalProperties": {
+                        "type": "object",
+                        "properties": {
+                            "type": {
+                                "type": "string",
+                                "enum": ["string", "integer", "number", "numeric"]
+                            },
+                            "description": {
+                                "type": "string"
+                            },
+                            "enum": {
+                                "type": "array",
+                                "items": {
+                                    "type": ["string", "number", "integer"]
+                                },
+                                "minItems": 1,
+                                "uniqueItems": true
+                            },
+                            "minimum": {
+                                "type": "number"
+                            }
+                        },
+                        "additionalProperties": false
+                    }
                 }
             },
             "required": ["contrasts"],

docs/usage.md

@@ -211,6 +211,22 @@ The necessary fields in order are:
 - `formula` - A string representation of the model formula. It is used to build the design matrix.
 - `make_contrasts_str` - An explicit literal contrast string (e.g., "treatmenthND6 - treatmentmCherry") that is passed directly to [`limma::makeContrasts()`](https://rdrr.io/bioc/limma/man/makeContrasts.html) in `VARIANCEPARTITION_DREAM`, `LIMMA_DIFFERENTIAL` and `DESEQ2_DIFFERENTIAL`. The parameter names must be syntactically valid variable names in R (see [`make.names`](https://stat.ethz.ch/R-manual/R-devel/library/base/html/make.names.html)). This field provides full control for complex designs. Requires `formula`.
 
+YAML contrast files can also include optional variable definitions. In the prototype implementation, `enum` values are used to set factor levels in the listed order before model matrices and contrast strings are evaluated:
+
+```yaml
+variables:
+  condition:
+    type: string
+    description: treatment group
+    enum: ["treated", "control"]
+contrasts:
+  - id: condition_treated_control
+    formula: "~ condition"
+    make_contrasts_str: "conditioncontrol"
+```
+
+Supported prototype fields are `type`, `description`, `enum`, and `minimum`. `enum` is the important field for categorical variables because its first value becomes the model baseline for intercept-style formulas.
+
 > [!IMPORTANT]
 >
 > - YAML contrast definitions using `comparison` **and** those using `formula` are both supported by all differential methods.