diff --git a/README.md b/README.md index 017eb02..e601bfa 100644 --- a/README.md +++ b/README.md @@ -1,6 +1,6 @@ # ood-proteinfold (OnDemand App) -`ood-proteinfold` is an [Open OnDemand](https://openondemand.org/) interactive app for computational protein structure prediction using [Proteinfold](https://nf-co.re/proteinfold/) ([AlphaFold2](https://github.com/google-deepmind/alphafold/), [Boltz](https://github.com/jwohlwend/boltz/), [ColabFold](https://github.com/sokrypton/ColabFold), [ESMFold](https://github.com/facebookresearch/esm), and [RoseTTAFold-All-Atom](https://github.com/baker-laboratory/RoseTTAFold-All-Atom/)) originally built for [UNSW's Katana HPC](https://www.unsw.edu.au/research/facilities-and-infrastructure/find-an-instrument/restech-instruments/katana). +`ood-proteinfold` is an [Open OnDemand](https://openondemand.org/) interactive app for computational protein structure prediction using [Proteinfold](https://nf-co.re/proteinfold/) ([AlphaFold2](https://github.com/google-deepmind/alphafold/), [Boltz](https://github.com/jwohlwend/boltz/), [ColabFold](https://github.com/sokrypton/ColabFold) and [ESMFold](https://github.com/facebookresearch/esm) originally built for [UNSW's Katana HPC](https://www.unsw.edu.au/research/facilities-and-infrastructure/find-an-instrument/restech-instruments/katana). ## Features diff --git a/completed.html.erb b/completed.html.erb index 130adc1..164d1bb 100644 --- a/completed.html.erb +++ b/completed.html.erb @@ -69,8 +69,44 @@ def sanitize_live_log(text) compact.join("\n").strip end -base_out_dir = (connect.respond_to?(:base_out_dir) && !connect.base_out_dir.blank?) ? connect.base_out_dir : "/srv/scratch/#{connect.user}/proteinfold_output" -results_url_base = (connect.respond_to?(:results_url_base) && !connect.results_url_base.blank?) ? connect.results_url_base : "/pun/sys/dashboard/files/fs" +def sanitize_output_log(text) + return nil if text.nil? + + lines = text.split("\n") + cleaned = [] + index = 0 + + while index < lines.length + line = lines[index] + + if line.start_with?("WARN: [PBSPRO] queue status cannot be fetched") + index += 1 + index += 1 while index < lines.length && !lines[index].strip.empty? + index += 1 if index < lines.length && lines[index].strip.empty? + next + end + + if line.strip == "Cleaning up..." + break + end + + if line.match?(/^=+$/) && index + 1 < lines.length && lines[index + 1].include?("Resource Usage on ") + break + end + + cleaned << line + index += 1 + end + + while cleaned.any? && cleaned.last.strip.empty? + cleaned.pop + end + + cleaned.join("\n").strip +end + +base_out_dir = (connect.respond_to?(:base_out_dir) && !connect.base_out_dir.blank?) ? connect.base_out_dir : portal_default_base_out_dir +results_url_base = (connect.respond_to?(:results_url_base) && !connect.results_url_base.blank?) ? connect.results_url_base : portal_default_results_url_base session_output_dir = (connect.respond_to?(:session_output_dir) && !connect.session_output_dir.blank?) ? connect.session_output_dir : nil log_dir = session_output_dir || File.join(base_out_dir, connect.run_dir) live_log_path = File.join(log_dir, "live_log.log") @@ -79,7 +115,11 @@ output_log_path = File.join(log_dir, "output.log") output_log_url = file_url(results_url_base, output_log_path) summary_log_url = File.file?(output_log_path) ? output_log_url : live_log_url summary_log_label = File.file?(output_log_path) ? "output.log" : "live_log.log" -log_text = sanitize_live_log(tail_text(live_log_path)) +log_text = if File.file?(output_log_path) + sanitize_output_log(tail_text(output_log_path, 16384, 40)) +else + sanitize_live_log(tail_text(live_log_path)) +end %>
Log: <%= summary_log_label %>
diff --git a/form.yml.erb b/form.yml.erb index beaf2db..a3bee01 100644 --- a/form.yml.erb +++ b/form.yml.erb @@ -54,14 +54,12 @@ attributes:
  • AlphaFold2.3 High accuracy, Slower - Paper
  • ColabFold High accuracy - Paper
  • ESMFold Medium/Low accuracy, Fastest; No evolutionary sequence calculations - Paper
    • -
  • RoseTTAFold-All-Atom Medium accuracy, Slower; Supports all atom modelling - Paper
    • options: - ["Boltz", "boltz", data-hide-prot_mode: true, data-hide-full_dbs: true, data-hide-colabfold_num_recycles: true, data-hide-esmfold_num_recycles: true] - ["AlphaFold2", "alphafold2", data-hide-colabfold_num_recycles: true, data-hide-esmfold_num_recycles: true, data-hide-boltz_use_potentials: true] - ["ColabFold", "colabfold", data-hide-full_dbs: true, data-hide-esmfold_num_recycles: true, data-hide-boltz_use_potentials: true] - ["ESMFold", "esmfold", data-hide-full_dbs: true, data-hide-colabfold_num_recycles: true, data-hide-boltz_use_potentials: true] - - ["RoseTTAFold-All-Atom", "rosettafold_all_atom", data-hide-prot_mode: true, data-hide-full_dbs: true, data-hide-colabfold_num_recycles: true, data-hide-esmfold_num_recycles: true, data-hide-boltz_use_potentials: true] display: true prot_mode: diff --git a/template/script.sh.erb b/template/script.sh.erb index 93a6504..1a38b56 100755 --- a/template/script.sh.erb +++ b/template/script.sh.erb @@ -112,10 +112,10 @@ echo "${MODE}" # Validate input file format compatibility with selected method af_method="<%= context.af_method %>" -# Check YAML/YML compatibility - only supported by Boltz and RoseTTAFold-All-Atom +# Check YAML/YML compatibility - only supported by Boltz if [ "${MODE}" == "FASTA_INPUT" ] && [[ "${user_input}" =~ \.(yaml|yml)$ ]]; then - if [ "${af_method}" != "boltz" ] && [ "${af_method}" != "rosettafold_all_atom" ]; then - echo "ERROR: YAML/YML input files are only supported by Boltz and RoseTTAFold-All-Atom methods." + if [ "${af_method}" != "boltz" ]; then + echo "ERROR: YAML/YML input files are only supported by Boltz." echo "Current method: ${af_method}" exit 1 fi @@ -133,10 +133,10 @@ fi # Check directory contents for incompatible file types if [ "${MODE}" == "DIRECTORY" ] && [ -d "${user_input}" ]; then # Check for YAML/YML files in directory - if [ "${af_method}" != "boltz" ] && [ "${af_method}" != "rosettafold_all_atom" ]; then + if [ "${af_method}" != "boltz" ]; then yaml_files=$(find "${user_input}" -type f \( -name "*.yaml" -o -name "*.yml" \) 2>/dev/null || true) if [ ! -z "${yaml_files}" ]; then - echo "ERROR: Directory contains YAML/YML files, which are only supported by Boltz and RoseTTAFold-All-Atom methods." + echo "ERROR: Directory contains YAML/YML files, which are only supported by Boltz." echo "Current method: ${af_method}" echo "Found YAML/YML files:" echo "${yaml_files}" @@ -286,9 +286,6 @@ if [ "<%= context.af_method %>" == "esmfold" ]; then fi fi -[[ "<%= context.af_method %>" == "rosettafold_all_atom" ]] && PROT_MODE="rosettafold_all_atom" && ARGS="${ARGS} --rosettafold_all_atom_db ${DB_PATH}" -[[ "<%= context.af_method %>" == "helixfold3" ]] && PROT_MODE="helixfold3" && ARGS="${ARGS} --helixfold3_db ${DB_PATH}" - nextflow -c "${PROJECT_ROOT}/kod_proteinfold-${RUN_ENVIRONMENT}.config" run ${REPOSITORY} -r ${BRANCH} -latest \ --input "${SAMPLESHEET}" \ --outdir "${OUT_DIR}" \