use sip hash algorithm [ci fast]

Signed-off-by: Paolo Di Tommaso <[email protected]>
nextflow-io · Nov 10, 2024 · 7edc0e7 · 7edc0e7
1 parent 76696f4
commit 7edc0e7
Show file tree

Hide file tree

Showing 2 changed files with 21 additions and 11 deletions.
diff --git a/modules/nextflow/src/main/groovy/nextflow/conda/CondaCache.groovy b/modules/nextflow/src/main/groovy/nextflow/conda/CondaCache.groovy
@@ -22,6 +22,7 @@ import java.nio.file.Path
 import java.nio.file.Paths
 import java.util.concurrent.ConcurrentHashMap
 
+import com.google.common.hash.Hashing
 import groovy.transform.CompileStatic
 import groovy.transform.PackageScope
 import groovy.util.logging.Slf4j
@@ -32,7 +33,6 @@ import nextflow.file.FileMutex
 import nextflow.util.CacheHelper
 import nextflow.util.Duration
 import nextflow.util.Escape
-import org.yaml.snakeyaml.Yaml
 /**
  * Handle Conda environment creation and caching
  *
@@ -166,6 +166,18 @@ class CondaCache {
         str.endsWith('.txt') && !str.contains('\n')
     }
 
+    static protected String sipHash(CharSequence data) {
+        Hashing
+            .sipHash24()
+            .newHasher()
+            .putUnencodedChars(data)
+            .hash()
+            .toString()
+    }
+
+    static protected String sipHash(Path path) {
+        sipHash(path.toAbsolutePath().normalize().toString())
+    }
 
     /**
      * Get the path on the file system where store a Conda environment
@@ -188,7 +200,7 @@ class CondaCache {
             try {
                 final path = condaEnv as Path
                 content = path.text
-                name = 'env-' + CacheHelper.hasher(path.toString()).hash().toString()
+                name = 'env-' + sipHash(path)
 
             }
             catch( NoSuchFileException e ) {
@@ -202,7 +214,7 @@ class CondaCache {
             try {
                 final path = condaEnv as Path
                 content = path.text
-                name = 'env-'+ CacheHelper.hasher(path.toString()).hash().toString()
+                name = 'env-' + sipHash(path)
             }
             catch( NoSuchFileException e ) {
                 throw new IllegalArgumentException("Conda environment file does not exist: $condaEnv")

diff --git a/modules/nextflow/src/test/groovy/nextflow/conda/CondaCacheTest.groovy b/modules/nextflow/src/test/groovy/nextflow/conda/CondaCacheTest.groovy
@@ -19,8 +19,6 @@ package nextflow.conda
 import java.nio.file.Files
 import java.nio.file.Paths
 
-import nextflow.util.CacheHelper
-
 import spock.lang.Specification
 /**
  *
@@ -91,7 +89,7 @@ class CondaCacheTest extends Specification {
         def cache = Spy(CondaCache)
         def BASE = Paths.get('/conda/envs')
         def ENV = folder.resolve('foo.yml')
-        def hash = CacheHelper.hasher(ENV.toString()).hash().toString()
+        def hash = CondaCache.sipHash(ENV)
         ENV.text = '''
             channels:
               - bioconda
@@ -107,7 +105,7 @@ class CondaCacheTest extends Specification {
         then:
         1 * cache.isYamlFilePath(ENV.toString())
         1 * cache.getCacheDir() >> BASE
-        prefix.toString() == "/conda/envs/env-$hash-9416240708c49c4e627414b46a743664"
+        prefix.toString() == "/conda/envs/env-${hash}-9416240708c49c4e627414b46a743664"
 
         cleanup:
         folder?.deleteDir()
@@ -120,7 +118,7 @@ class CondaCacheTest extends Specification {
         def cache = Spy(CondaCache)
         def BASE = Paths.get('/conda/envs')
         def ENV = Files.createTempFile('test','.yml')
-        def hash = CacheHelper.hasher(ENV.toString()).hash().toString()
+        def hash = CondaCache.sipHash(ENV)
         ENV.text = '''  
             name: my-env-1.1
             channels:
@@ -138,7 +136,7 @@ class CondaCacheTest extends Specification {
         then:
         1 * cache.isYamlFilePath(ENV.toString())
         1 * cache.getCacheDir() >> BASE
-        prefix.toString() == "/conda/envs/env-$hash-e7fafe40ca966397a2c0d9bed7181aa7"
+        prefix.toString() == "/conda/envs/env-${hash}-e7fafe40ca966397a2c0d9bed7181aa7"
 
     }
 
@@ -149,7 +147,7 @@ class CondaCacheTest extends Specification {
         def cache = Spy(CondaCache)
         def BASE = Paths.get('/conda/envs')
         def ENV = folder.resolve('bar.txt')
-        def hash = CacheHelper.hasher(ENV.toString()).hash().toString()
+        def hash = CondaCache.sipHash(ENV)
         ENV.text = '''
                 star=2.5.4a
                 bwa=0.7.15   
@@ -163,7 +161,7 @@ class CondaCacheTest extends Specification {
         1 * cache.isYamlFilePath(ENV.toString())
         1 * cache.isTextFilePath(ENV.toString())
         1 * cache.getCacheDir() >> BASE
-        prefix.toString() == "/conda/envs/env-$hash-8a4aa7db8ddb8ce4eb4d450d4814a437"
+        prefix.toString() == "/conda/envs/env-${hash}-8a4aa7db8ddb8ce4eb4d450d4814a437"
 
         cleanup:
         folder?.deleteDir()