Merge branch 'mod_models_v201' into 'master'

mod_models: released updated v2.0.1 models to correct error in motif specification in config See merge request machine-learning/dorado!1215
nanoporetech · Oct 3, 2024 · f74d891 · f74d891
2 parents 5d68cc4 + e83c494
commit f74d891
Show file tree

Hide file tree

Showing 6 changed files with 22 additions and 16 deletions.
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
@@ -527,6 +527,7 @@ build_archive:macos:x64:
     MODEL_SPEED: hac
     MODEL_VERSION: v4.2.0
     NO_TEST_DORADO_CORRECT: 0
+    NO_TEST_DUPLEX: 0
   script:
     # Do everything inside a newly created folder, since some runners only run
     # tests and hence never perform a git clean.
@@ -600,6 +601,7 @@ test:linux:arm64:bionic:
     MODEL_SPEED: hac
     MODEL_VERSION: v4.2.0
     NO_TEST_DORADO_CORRECT: 1
+    NO_TEST_DUPLEX: 1
   needs:
     - build:linux:arm64:bionic
 
@@ -631,6 +633,8 @@ test:macos:x64:
     BATCH: 0
     MODEL_SPEED: hac
     MODEL_VERSION: v4.2.0
+    NO_TEST_DORADO_CORRECT: 1
+    NO_TEST_DUPLEX: 1
   timeout: 1h 30m
 
 # Test that you can run dorado in a clean cuda 20.04 environment

diff --git a/README.md b/README.md
@@ -388,7 +388,7 @@ The versioning of modification models is bound to the basecalling model. This me
 | :-------- | :------- | :--- | :--- |
 | **dna_r10.4.1_e8.2_400bps_fast@v5.0.0** | | | 5 kHz |
 | **dna_r10.4.1_e8.2_400bps_hac@v5.0.0** | 4mC_5mC<br />5mCG_5hmCG<br />5mC_5hmC<br />6mA<br /> | v2<br />v2<br />v2<br />v2 | 5 kHz |
-| **dna_r10.4.1_e8.2_400bps_sup@v5.0.0** | 4mC_5mC<br />5mCG_5hmCG<br />5mC_5hmC<br />6mA<br /> | v2<br />v2<br />v2<br />v2 | 5 kHz |
+| **dna_r10.4.1_e8.2_400bps_sup@v5.0.0** | 4mC_5mC<br />5mCG_5hmCG<br />5mC_5hmC<br />6mA<br /> | v2<br />v2.0.1<br />v2.0.1<br />v2 | 5 kHz |
 | dna_r10.4.1_e8.2_400bps_fast@v4.3.0 | | | 5 kHz |
 | dna_r10.4.1_e8.2_400bps_hac@v4.3.0 | 5mCG_5hmCG<br />5mC_5hmC<br />6mA<br /> | v1<br />v1<br />v2 | 5 kHz |
 | dna_r10.4.1_e8.2_400bps_sup@v4.3.0 | 5mCG_5hmCG<br />5mC_5hmC<br />6mA<br /> | v1<br />v1<br />v2 | 5 kHz |

diff --git a/dorado/models/metadata.cpp b/dorado/models/metadata.cpp
@@ -49,14 +49,14 @@ const std::unordered_map<std::string, ModelVersion> map = {
         {"v0.0.0", ModelVersion::v0_0_0}, {"v0.1.0", ModelVersion::v0_1_0},
         {"v1.0.0", ModelVersion::v1_0_0}, {"v1.1.0", ModelVersion::v1_1_0},
         {"v1.2.0", ModelVersion::v1_2_0}, {"v2.0.0", ModelVersion::v2_0_0},
-        {"v3.0.0", ModelVersion::v3_0_0}, {"v3.0.1", ModelVersion::v3_0_1},
-        {"v3.1.0", ModelVersion::v3_1_0}, {"v3.3.0", ModelVersion::v3_3_0},
-        {"v3.4.0", ModelVersion::v3_4_0}, {"v3.5.0", ModelVersion::v3_5_0},
-        {"v3.5.2", ModelVersion::v3_5_2}, {"v3.6.0", ModelVersion::v3_6_0},
-        {"v4.0.0", ModelVersion::v4_0_0}, {"v4.1.0", ModelVersion::v4_1_0},
-        {"v4.2.0", ModelVersion::v4_2_0}, {"v4.3.0", ModelVersion::v4_3_0},
-        {"v5.0.0", ModelVersion::v5_0_0}, {"v5.1.0", ModelVersion::v5_1_0},
-        {"latest", ModelVersion::NONE}};
+        {"v2.0.1", ModelVersion::v2_0_1}, {"v3.0.0", ModelVersion::v3_0_0},
+        {"v3.0.1", ModelVersion::v3_0_1}, {"v3.1.0", ModelVersion::v3_1_0},
+        {"v3.3.0", ModelVersion::v3_3_0}, {"v3.4.0", ModelVersion::v3_4_0},
+        {"v3.5.0", ModelVersion::v3_5_0}, {"v3.5.2", ModelVersion::v3_5_2},
+        {"v3.6.0", ModelVersion::v3_6_0}, {"v4.0.0", ModelVersion::v4_0_0},
+        {"v4.1.0", ModelVersion::v4_1_0}, {"v4.2.0", ModelVersion::v4_2_0},
+        {"v4.3.0", ModelVersion::v4_3_0}, {"v5.0.0", ModelVersion::v5_0_0},
+        {"v5.1.0", ModelVersion::v5_1_0}, {"latest", ModelVersion::NONE}};
 }  // namespace version
 
 const std::unordered_map<std::string, ModelVariant>& model_variants_map() {

diff --git a/dorado/models/metadata.h b/dorado/models/metadata.h
@@ -42,6 +42,7 @@ enum class ModelVersion : uint8_t {
     v1_1_0,
     v1_2_0,
     v2_0_0,
+    v2_0_1,
     v3_0_0,
     v3_0_1,
     v3_1_0,

diff --git a/dorado/models/models.cpp b/dorado/models/models.cpp
@@ -863,11 +863,11 @@ const std::vector<ModelInfo> models = {
                 ModsVariantPair{ModsVariant::M_5mC_5hmC, VV::v2_0_0},
         },
         ModelInfo{
-                "dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mC_5hmC@v2",
-                "36de2e58edaf1e1a53bca0ebf029164112b9dbaad413672dde45efb093b7fcf6",
+                "dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mC_5hmC@v2.0.1",
+                "757dabc280e25f1c442fcfeb3e1f4d44a2d445e0ea89bb30c15e4757879111be",
                 CC::DNA_R10_4_1_E8_2_400BPS_5KHZ,
                 ModelVariantPair{ModelVariant::SUP, VV::v5_0_0},
-                ModsVariantPair{ModsVariant::M_5mC_5hmC, VV::v2_0_0},
+                ModsVariantPair{ModsVariant::M_5mC_5hmC, VV::v2_0_1},
         },
         // 5mC+5hmC CG-context HAC and SUP
         ModelInfo{
@@ -892,11 +892,11 @@ const std::vector<ModelInfo> models = {
                 ModsVariantPair{ModsVariant::M_5mCG_5hmCG, VV::v2_0_0},
         },
         ModelInfo{
-                "dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mCG_5hmCG@v2",
-                "2e2d4de2ec1df90b37c50b3367bea90f7b9dfab11b90e98ee6963876589be4cc",
+                "dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mCG_5hmCG@v2.0.1",
+                "c8ebafd13008a919232cd45514e07ea929509a5e20254c73b9eff2cd0e5a4786",
                 CC::DNA_R10_4_1_E8_2_400BPS_5KHZ,
                 ModelVariantPair{ModelVariant::SUP, VV::v5_0_0},
-                ModsVariantPair{ModsVariant::M_5mCG_5hmCG, VV::v2_0_0},
+                ModsVariantPair{ModsVariant::M_5mCG_5hmCG, VV::v2_0_1},
         },
         // 6mA all-context HAC and SUP
         ModelInfo{

diff --git a/tests/test_simple_basecaller_execution.sh b/tests/test_simple_basecaller_execution.sh
@@ -188,7 +188,8 @@ dorado_aligner_options_test() (
 )
 dorado_aligner_options_test
 
-if ! uname -r | grep -q -E 'tegra|minit'; then
+# Skip duplex tests if NO_TEST_DUPLEX is set.
+if [[ "${NO_TEST_DUPLEX}" -ne "1" ]]; then
     echo dorado duplex basespace test stage
     $dorado_bin duplex basespace $data_dir/basespace/pairs.bam --threads 1 --pairs $data_dir/basespace/pairs.txt > $output_dir/calls.bam
-Original file line number
+Diff line change
@@ Expand Up / @@ -42,6 +42,7 @@ enum class ModelVersion : uint8_t { @@
         v1_1_0,
         v1_2_0,
         v2_0_0,
+        v2_0_1,
         v3_0_0,
         v3_0_1,
         v3_1_0,
@@ Expand Down @@