Merge pull request #211 from VEuPathDB/issue-202

Validate dash count on first line of each sequence.
VEuPathDB · Dec 14, 2022 · 32df5fd · 32df5fd
2 parents 708603d + 687e7b1
commit 32df5fd
Show file tree

Hide file tree

Showing 2 changed files with 56 additions and 0 deletions.
diff --git a/rest-service/src/main/kotlin/mb/api/service/valid/SequenceValidationError.kt b/rest-service/src/main/kotlin/mb/api/service/valid/SequenceValidationError.kt
@@ -25,4 +25,9 @@ class SequenceLengthValidationError(
 class SequenceEmptyValidationError : SequenceValidationError {
   override val message: String
     get() = "Empty query.  Query must have 1 or more sequences."
+}
+
+class SequenceDashesValidationError(val sequence: Int) : SequenceValidationError {
+  override val message: String
+    get() = "The first line of input sequence #$sequence contained too many dashes."
 }
diff --git a/rest-service/src/main/kotlin/mb/api/service/valid/SequenceValidator.kt b/rest-service/src/main/kotlin/mb/api/service/valid/SequenceValidator.kt
@@ -2,6 +2,7 @@ package mb.api.service.valid
 
 import org.veupathdb.lib.blast.BlastTool
 import java.util.*
+import kotlin.math.min
 
 interface SequenceValidator {
   /**
@@ -32,6 +33,8 @@ interface SequenceValidator {
    * @return Whether all characters in the input {@code CharSequence} are valid.
    */
   fun validate(sequence: Int, seq: String): SequenceValidationError? {
+    validateFirstLine(sequence, seq)?.let { return it }
+
     // Input scanner
     val scan = Scanner(seq)
     // Current line number
@@ -65,6 +68,54 @@ interface SequenceValidator {
     return null
   }
 
+  /**
+   * Validates the first line in a sequence to ensure that it passes the first
+   * line check performed by the BLAST+ CLI tool:
+   * ```
+   * if (bad >= good / 3  &&
+   *   (len_to_check > 3  ||  good == 0  ||  bad > good))
+   * {
+   *   FASTA_ERROR( LineNumber(),
+   *     "CFastaReader: Near line " << LineNumber()
+   *     << ", there's a line that doesn't look like plausible data, "
+   *     "but it's not marked as defline or comment.",
+   *     CObjReaderParseException::eFormat);
+   * }
+   * ```
+   */
+  private fun validateFirstLine(sequence: Int, seq: String): SequenceValidationError? {
+    val scan = Scanner(seq)
+
+    while (scan.hasNextLine()) {
+      val line = scan.nextLine()
+
+      if (line.isNotBlank() && line[0] != '>') {
+        var dashes    = 0
+        var nonDashes = 0
+
+        // Count the dashes + nondashes in the line
+        for (i in 0 .. min(line.length, 70)) {
+          when (line[i]) {
+            '-'  -> dashes++
+            else -> nonDashes++
+          }
+        }
+
+        val checked = dashes+nonDashes
+
+        return if (
+          dashes >= nonDashes / 3
+          && (checked > 3 || nonDashes == 0 || dashes > nonDashes)
+        )
+          SequenceDashesValidationError(sequence)
+        else
+          null
+      }
+    }
+
+    return SequenceEmptyValidationError()
+  }
+
   companion object {
     fun getValidator(tool: BlastTool): SequenceValidator {
       return when (tool) {