From ab7055db6fac4f4c5f7ed51a5779fa08c995b136 Mon Sep 17 00:00:00 2001 From: Tom Harrop Date: Tue, 11 Jun 2024 13:08:53 +1000 Subject: [PATCH] working LOC file --- tools/dorado/README.md | 30 ++++++++++ tools/dorado/dorado.xml | 13 ++--- tools/dorado/macros.xml | 5 +- tools/dorado/test-data/bowtie_indices.loc | 1 - tools/dorado/test-data/dorado_models.loc | 1 + .../dorado/tool-data/dorado_models.loc.sample | 58 +++++++++++++++++++ tools/dorado/tool_data_table_conf.xml.sample | 7 +-- tools/dorado/tool_data_table_conf.xml.test | 7 +-- 8 files changed, 103 insertions(+), 19 deletions(-) delete mode 120000 tools/dorado/test-data/bowtie_indices.loc create mode 120000 tools/dorado/test-data/dorado_models.loc create mode 100644 tools/dorado/tool-data/dorado_models.loc.sample diff --git a/tools/dorado/README.md b/tools/dorado/README.md index e69de29b..28be521b 100644 --- a/tools/dorado/README.md +++ b/tools/dorado/README.md @@ -0,0 +1,30 @@ + +## Tool versions + +Dorado is distributed on +[DockerHub](https://hub.docker.com/r/nanoporetech/dorado/tags) by nanoporetech, +but not tagged with a version. + +That means the hash for the current version has to be hard-coded into the +wrapper. Unfortunately you have to pull a >6 GB container just to check the tool +version. At least you can update the list of models at the same time (see +below). + +**Make sure you do this when you update the wrapper**! + +## Basecalling models + +The models are bundled in the container at `/models` and made available by the +`dorado_models.loc` file. To update the list, modify +`tool-data/dorado_models.loc.sample`. Note that if ONT remove models from the +container, doing this will also make them unavailable to Galaxy. Check the diff +before you merge. + +Here's a one-liner to **replace** the contents of the loc file with the models that are bundled in the container `nanoporetech/dorado:shac2d8bc91ca2d043fed84d06cca92aaeb62bcc1cd`. + +```bash +apptainer exec docker://nanoporetech/dorado:shac2d8bc91ca2d043fed84d06cca92aaeb62bcc1cd \ + ls /models | \ + awk '{print $0 "\t" $0 "\t/models/" $0}' \ + > tool-data/dorado_models.loc.sample +``` \ No newline at end of file diff --git a/tools/dorado/dorado.xml b/tools/dorado/dorado.xml index f8051dc4..f959ca42 100644 --- a/tools/dorado/dorado.xml +++ b/tools/dorado/dorado.xml @@ -1,4 +1,4 @@ - + macros.xml @@ -10,19 +10,16 @@ ln -s '$pod5_file' ./reads.pod5 && dorado basecaller --emit-sam -## TODO: get a list of models from /models (bundled with the container) -"/models/dna_r9.4.1_e8_fast@v3.4" +'${model.fields.path}' reads.pod5 > output.sam ]]> - - - + + + diff --git a/tools/dorado/macros.xml b/tools/dorado/macros.xml index 2eddfc86..0a4a37ab 100644 --- a/tools/dorado/macros.xml +++ b/tools/dorado/macros.xml @@ -1,6 +1,7 @@ - - c2d8bc91ca2d043fed84d06cca92aaeb62bcc1cd + + 0.7.1 + 1c65eb070a9fc1d88710c4dc09b06541f96fdd28 nanoporetech/dorado:sha@CONTAINER_HASH@ diff --git a/tools/dorado/test-data/bowtie_indices.loc b/tools/dorado/test-data/bowtie_indices.loc deleted file mode 120000 index f8185c61..00000000 --- a/tools/dorado/test-data/bowtie_indices.loc +++ /dev/null @@ -1 +0,0 @@ -../tool-data/bowtie_indices.loc.sample \ No newline at end of file diff --git a/tools/dorado/test-data/dorado_models.loc b/tools/dorado/test-data/dorado_models.loc new file mode 120000 index 00000000..fb6fc000 --- /dev/null +++ b/tools/dorado/test-data/dorado_models.loc @@ -0,0 +1 @@ +../tool-data/dorado_models.loc.sample \ No newline at end of file diff --git a/tools/dorado/tool-data/dorado_models.loc.sample b/tools/dorado/tool-data/dorado_models.loc.sample new file mode 100644 index 00000000..e16e5685 --- /dev/null +++ b/tools/dorado/tool-data/dorado_models.loc.sample @@ -0,0 +1,58 @@ +dna_r10.4.1_e8.2_260bps_fast@v4.1.0 dna_r10.4.1_e8.2_260bps_fast@v4.1.0 /models/dna_r10.4.1_e8.2_260bps_fast@v4.1.0 +dna_r10.4.1_e8.2_260bps_fast@v4.1.0_5mCG_5hmCG@v2 dna_r10.4.1_e8.2_260bps_fast@v4.1.0_5mCG_5hmCG@v2 /models/dna_r10.4.1_e8.2_260bps_fast@v4.1.0_5mCG_5hmCG@v2 +dna_r10.4.1_e8.2_260bps_hac@v4.1.0 dna_r10.4.1_e8.2_260bps_hac@v4.1.0 /models/dna_r10.4.1_e8.2_260bps_hac@v4.1.0 +dna_r10.4.1_e8.2_260bps_hac@v4.1.0_5mCG_5hmCG@v2 dna_r10.4.1_e8.2_260bps_hac@v4.1.0_5mCG_5hmCG@v2 /models/dna_r10.4.1_e8.2_260bps_hac@v4.1.0_5mCG_5hmCG@v2 +dna_r10.4.1_e8.2_260bps_sup@v4.1.0 dna_r10.4.1_e8.2_260bps_sup@v4.1.0 /models/dna_r10.4.1_e8.2_260bps_sup@v4.1.0 +dna_r10.4.1_e8.2_260bps_sup@v4.1.0_5mCG_5hmCG@v2 dna_r10.4.1_e8.2_260bps_sup@v4.1.0_5mCG_5hmCG@v2 /models/dna_r10.4.1_e8.2_260bps_sup@v4.1.0_5mCG_5hmCG@v2 +dna_r10.4.1_e8.2_400bps_fast@v4.1.0 dna_r10.4.1_e8.2_400bps_fast@v4.1.0 /models/dna_r10.4.1_e8.2_400bps_fast@v4.1.0 +dna_r10.4.1_e8.2_400bps_fast@v4.1.0_5mCG_5hmCG@v2 dna_r10.4.1_e8.2_400bps_fast@v4.1.0_5mCG_5hmCG@v2 /models/dna_r10.4.1_e8.2_400bps_fast@v4.1.0_5mCG_5hmCG@v2 +dna_r10.4.1_e8.2_400bps_fast@v4.2.0 dna_r10.4.1_e8.2_400bps_fast@v4.2.0 /models/dna_r10.4.1_e8.2_400bps_fast@v4.2.0 +dna_r10.4.1_e8.2_400bps_fast@v4.2.0_5mCG_5hmCG@v2 dna_r10.4.1_e8.2_400bps_fast@v4.2.0_5mCG_5hmCG@v2 /models/dna_r10.4.1_e8.2_400bps_fast@v4.2.0_5mCG_5hmCG@v2 +dna_r10.4.1_e8.2_400bps_fast@v4.3.0 dna_r10.4.1_e8.2_400bps_fast@v4.3.0 /models/dna_r10.4.1_e8.2_400bps_fast@v4.3.0 +dna_r10.4.1_e8.2_400bps_fast@v5.0.0 dna_r10.4.1_e8.2_400bps_fast@v5.0.0 /models/dna_r10.4.1_e8.2_400bps_fast@v5.0.0 +dna_r10.4.1_e8.2_400bps_hac@v4.1.0 dna_r10.4.1_e8.2_400bps_hac@v4.1.0 /models/dna_r10.4.1_e8.2_400bps_hac@v4.1.0 +dna_r10.4.1_e8.2_400bps_hac@v4.1.0_5mCG_5hmCG@v2 dna_r10.4.1_e8.2_400bps_hac@v4.1.0_5mCG_5hmCG@v2 /models/dna_r10.4.1_e8.2_400bps_hac@v4.1.0_5mCG_5hmCG@v2 +dna_r10.4.1_e8.2_400bps_hac@v4.3.0 dna_r10.4.1_e8.2_400bps_hac@v4.3.0 /models/dna_r10.4.1_e8.2_400bps_hac@v4.3.0 +dna_r10.4.1_e8.2_400bps_hac@v4.3.0_5mCG_5hmCG@v1 dna_r10.4.1_e8.2_400bps_hac@v4.3.0_5mCG_5hmCG@v1 /models/dna_r10.4.1_e8.2_400bps_hac@v4.3.0_5mCG_5hmCG@v1 +dna_r10.4.1_e8.2_400bps_hac@v4.3.0_5mC_5hmC@v1 dna_r10.4.1_e8.2_400bps_hac@v4.3.0_5mC_5hmC@v1 /models/dna_r10.4.1_e8.2_400bps_hac@v4.3.0_5mC_5hmC@v1 +dna_r10.4.1_e8.2_400bps_hac@v4.3.0_6mA@v2 dna_r10.4.1_e8.2_400bps_hac@v4.3.0_6mA@v2 /models/dna_r10.4.1_e8.2_400bps_hac@v4.3.0_6mA@v2 +dna_r10.4.1_e8.2_400bps_hac@v5.0.0 dna_r10.4.1_e8.2_400bps_hac@v5.0.0 /models/dna_r10.4.1_e8.2_400bps_hac@v5.0.0 +dna_r10.4.1_e8.2_400bps_hac@v5.0.0_4mC_5mC@v1 dna_r10.4.1_e8.2_400bps_hac@v5.0.0_4mC_5mC@v1 /models/dna_r10.4.1_e8.2_400bps_hac@v5.0.0_4mC_5mC@v1 +dna_r10.4.1_e8.2_400bps_hac@v5.0.0_5mCG_5hmCG@v1 dna_r10.4.1_e8.2_400bps_hac@v5.0.0_5mCG_5hmCG@v1 /models/dna_r10.4.1_e8.2_400bps_hac@v5.0.0_5mCG_5hmCG@v1 +dna_r10.4.1_e8.2_400bps_hac@v5.0.0_5mC_5hmC@v1 dna_r10.4.1_e8.2_400bps_hac@v5.0.0_5mC_5hmC@v1 /models/dna_r10.4.1_e8.2_400bps_hac@v5.0.0_5mC_5hmC@v1 +dna_r10.4.1_e8.2_400bps_hac@v5.0.0_6mA@v1 dna_r10.4.1_e8.2_400bps_hac@v5.0.0_6mA@v1 /models/dna_r10.4.1_e8.2_400bps_hac@v5.0.0_6mA@v1 +dna_r10.4.1_e8.2_400bps_sup@v4.1.0 dna_r10.4.1_e8.2_400bps_sup@v4.1.0 /models/dna_r10.4.1_e8.2_400bps_sup@v4.1.0 +dna_r10.4.1_e8.2_400bps_sup@v4.1.0_5mCG_5hmCG@v2 dna_r10.4.1_e8.2_400bps_sup@v4.1.0_5mCG_5hmCG@v2 /models/dna_r10.4.1_e8.2_400bps_sup@v4.1.0_5mCG_5hmCG@v2 +dna_r10.4.1_e8.2_400bps_sup@v4.3.0 dna_r10.4.1_e8.2_400bps_sup@v4.3.0 /models/dna_r10.4.1_e8.2_400bps_sup@v4.3.0 +dna_r10.4.1_e8.2_400bps_sup@v4.3.0_5mCG_5hmCG@v1 dna_r10.4.1_e8.2_400bps_sup@v4.3.0_5mCG_5hmCG@v1 /models/dna_r10.4.1_e8.2_400bps_sup@v4.3.0_5mCG_5hmCG@v1 +dna_r10.4.1_e8.2_400bps_sup@v4.3.0_5mC_5hmC@v1 dna_r10.4.1_e8.2_400bps_sup@v4.3.0_5mC_5hmC@v1 /models/dna_r10.4.1_e8.2_400bps_sup@v4.3.0_5mC_5hmC@v1 +dna_r10.4.1_e8.2_400bps_sup@v4.3.0_6mA@v2 dna_r10.4.1_e8.2_400bps_sup@v4.3.0_6mA@v2 /models/dna_r10.4.1_e8.2_400bps_sup@v4.3.0_6mA@v2 +dna_r10.4.1_e8.2_400bps_sup@v5.0.0 dna_r10.4.1_e8.2_400bps_sup@v5.0.0 /models/dna_r10.4.1_e8.2_400bps_sup@v5.0.0 +dna_r10.4.1_e8.2_400bps_sup@v5.0.0_4mC_5mC@v1 dna_r10.4.1_e8.2_400bps_sup@v5.0.0_4mC_5mC@v1 /models/dna_r10.4.1_e8.2_400bps_sup@v5.0.0_4mC_5mC@v1 +dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mCG_5hmCG@v1 dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mCG_5hmCG@v1 /models/dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mCG_5hmCG@v1 +dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mC_5hmC@v1 dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mC_5hmC@v1 /models/dna_r10.4.1_e8.2_400bps_sup@v5.0.0_5mC_5hmC@v1 +dna_r10.4.1_e8.2_400bps_sup@v5.0.0_6mA@v1 dna_r10.4.1_e8.2_400bps_sup@v5.0.0_6mA@v1 /models/dna_r10.4.1_e8.2_400bps_sup@v5.0.0_6mA@v1 +dna_r10.4.1_e8.2_5khz_stereo@v1.3 dna_r10.4.1_e8.2_5khz_stereo@v1.3 /models/dna_r10.4.1_e8.2_5khz_stereo@v1.3 +dna_r9.4.1_e8_fast@v3.4 dna_r9.4.1_e8_fast@v3.4 /models/dna_r9.4.1_e8_fast@v3.4 +dna_r9.4.1_e8_fast@v3.4_5mCG@v0.1 dna_r9.4.1_e8_fast@v3.4_5mCG@v0.1 /models/dna_r9.4.1_e8_fast@v3.4_5mCG@v0.1 +dna_r9.4.1_e8_fast@v3.4_5mCG_5hmCG@v0 dna_r9.4.1_e8_fast@v3.4_5mCG_5hmCG@v0 /models/dna_r9.4.1_e8_fast@v3.4_5mCG_5hmCG@v0 +dna_r9.4.1_e8_hac@v3.3 dna_r9.4.1_e8_hac@v3.3 /models/dna_r9.4.1_e8_hac@v3.3 +dna_r9.4.1_e8_hac@v3.3_5mCG@v0.1 dna_r9.4.1_e8_hac@v3.3_5mCG@v0.1 /models/dna_r9.4.1_e8_hac@v3.3_5mCG@v0.1 +dna_r9.4.1_e8_hac@v3.3_5mCG_5hmCG@v0 dna_r9.4.1_e8_hac@v3.3_5mCG_5hmCG@v0 /models/dna_r9.4.1_e8_hac@v3.3_5mCG_5hmCG@v0 +dna_r9.4.1_e8_sup@v3.3 dna_r9.4.1_e8_sup@v3.3 /models/dna_r9.4.1_e8_sup@v3.3 +dna_r9.4.1_e8_sup@v3.3_5mCG@v0.1 dna_r9.4.1_e8_sup@v3.3_5mCG@v0.1 /models/dna_r9.4.1_e8_sup@v3.3_5mCG@v0.1 +dna_r9.4.1_e8_sup@v3.3_5mCG_5hmCG@v0 dna_r9.4.1_e8_sup@v3.3_5mCG_5hmCG@v0 /models/dna_r9.4.1_e8_sup@v3.3_5mCG_5hmCG@v0 +dna_r9.4.1_e8_sup@v3.6 dna_r9.4.1_e8_sup@v3.6 /models/dna_r9.4.1_e8_sup@v3.6 +rna002_70bps_fast@v3 rna002_70bps_fast@v3 /models/rna002_70bps_fast@v3 +rna002_70bps_hac@v3 rna002_70bps_hac@v3 /models/rna002_70bps_hac@v3 +rna004_130bps_fast@v3.0.1 rna004_130bps_fast@v3.0.1 /models/rna004_130bps_fast@v3.0.1 +rna004_130bps_fast@v5.0.0 rna004_130bps_fast@v5.0.0 /models/rna004_130bps_fast@v5.0.0 +rna004_130bps_hac@v3.0.1 rna004_130bps_hac@v3.0.1 /models/rna004_130bps_hac@v3.0.1 +rna004_130bps_hac@v5.0.0 rna004_130bps_hac@v5.0.0 /models/rna004_130bps_hac@v5.0.0 +rna004_130bps_hac@v5.0.0_m6A@v1 rna004_130bps_hac@v5.0.0_m6A@v1 /models/rna004_130bps_hac@v5.0.0_m6A@v1 +rna004_130bps_hac@v5.0.0_pseU@v1 rna004_130bps_hac@v5.0.0_pseU@v1 /models/rna004_130bps_hac@v5.0.0_pseU@v1 +rna004_130bps_sup@v3.0.1 rna004_130bps_sup@v3.0.1 /models/rna004_130bps_sup@v3.0.1 +rna004_130bps_sup@v3.0.1_m6A_DRACH@v1 rna004_130bps_sup@v3.0.1_m6A_DRACH@v1 /models/rna004_130bps_sup@v3.0.1_m6A_DRACH@v1 +rna004_130bps_sup@v5.0.0 rna004_130bps_sup@v5.0.0 /models/rna004_130bps_sup@v5.0.0 +rna004_130bps_sup@v5.0.0_m6A@v1 rna004_130bps_sup@v5.0.0_m6A@v1 /models/rna004_130bps_sup@v5.0.0_m6A@v1 +rna004_130bps_sup@v5.0.0_pseU@v1 rna004_130bps_sup@v5.0.0_pseU@v1 /models/rna004_130bps_sup@v5.0.0_pseU@v1 diff --git a/tools/dorado/tool_data_table_conf.xml.sample b/tools/dorado/tool_data_table_conf.xml.sample index c1d3e6f6..d766a72f 100644 --- a/tools/dorado/tool_data_table_conf.xml.sample +++ b/tools/dorado/tool_data_table_conf.xml.sample @@ -1,7 +1,6 @@ - - - value, dbkey, name, path - +
+ value, name, path +
\ No newline at end of file diff --git a/tools/dorado/tool_data_table_conf.xml.test b/tools/dorado/tool_data_table_conf.xml.test index 34f4fa4c..049af265 100644 --- a/tools/dorado/tool_data_table_conf.xml.test +++ b/tools/dorado/tool_data_table_conf.xml.test @@ -1,7 +1,6 @@ - - - value, dbkey, name, path - +
+ value, name, path +
\ No newline at end of file