braker almost readywip-braker

2 files changed, 324 insertions, 78 deletions

diff --git a/gn/packages/bioinformatics.scm b/gn/packages/bioinformatics.scm
index 677de1d..637027c 100644
--- a/gn/packages/bioinformatics.scm
+++ b/gn/packages/bioinformatics.scm
@@ -18,6 +18,7 @@
   #:use-module (guix build-system waf)
   #:use-module (gnu packages)
   #:use-module (gn packages crates-io)
+  #:use-module (gn packages databases)
   #:use-module (gn packages java)
   #:use-module (gn packages perl)
   #:use-module (gn packages python)
@@ -45,6 +46,7 @@
   #:use-module (gnu packages guile)
   #:use-module (gnu packages image)
   #:use-module (gnu packages imagemagick)
+  #:use-module (gnu packages java)
   #:use-module (gnu packages jemalloc)
   #:use-module (gnu packages linux)
   #:use-module (gnu packages machine-learning)
@@ -66,6 +68,7 @@
   #:use-module (gnu packages serialization)
   #:use-module (gnu packages shells)
   #:use-module (gnu packages statistics)
+  #:use-module (gnu packages sqlite)
   #:use-module (gnu packages tcl)
   #:use-module (gnu packages time)
   #:use-module (gnu packages tls)
@@ -2427,68 +2430,62 @@ To run the bundled rtg-tools software you will also need java.  The
     (build-system perl-build-system)
     (arguments
      `(#:modules ((srfi srfi-26)
-                  (guix build perl-build-system)
-                  (guix build utils))
+                  ,@%perl-build-system-modules)
        #:phases
        (modify-phases %standard-phases
-         (delete 'configure)
+         (delete 'configure)            ; No configure script.
          (delete 'build)
          (replace 'check
            (lambda* (#:key tests? #:allow-other-keys)
              (when tests?
-               (invoke "prove" "-l" "t"))))
+               (invoke "prove" "--verbose" "--lib" "t"))))
          (replace 'install
            (lambda* (#:key inputs outputs #:allow-other-keys)
-             (let* ((out (assoc-ref outputs "out"))
-                    (bin (string-append out "/bin")))
-               (with-directory-excursion "scripts"
-                 (for-each (cut install-file <> bin)
-                           (find-files "." "\\.(py|pl|pm)$"))
-                 (for-each
-                   (cut wrap-script <>
-                        `("PERL5LIB" ":" prefix (,(getenv "PERL5LIB")
-                                                 ,bin)))
-                   ;; braker.pl is the entry point so wrap it separately.
-                   (delete (string-append bin "/braker.pl")
-                           (find-files bin "\\.pl$")))
-                 (wrap-script (string-append bin "/braker.pl")
-                   `("PERL5LIB" ":" prefix (,(getenv "PERL5LIB")
-                                             ,bin))
-                   `("PATH" ":" prefix (;,(assoc-ref inputs "augustus")
-                                        ;,(assoc-ref inputs "genemark")
-                                        ,(assoc-ref inputs "bamtools")
-                                        ,(assoc-ref inputs "samtools")
-                                        ,(assoc-ref inputs "prothint")
-                                        ;,(assoc-ref inputs "genomethreader")
-                                        ,(assoc-ref inputs "spaln")
-                                        ,(assoc-ref inputs "exonerate")
-                                        ,(assoc-ref inputs "ncbi-blast")
-                                        ,(assoc-ref inputs "diamond")
-                                        ,(assoc-ref inputs "cdbfasta")
-                                        ;,(assoc-ref inputs "gushr")
-                                        ;,(assoc-ref inputs "ucsc")
-                                        ))
-                   ;`("GENEMARK_PATH" "=" (,(string-append (assoc-ref inputs "genemark") "/bin")))
-                   ;`("AUGUSTUS_BIN_PATH" "=" (,(string-append (assoc-ref inputs "augustus") "/bin")))
-                   ;`("AUGUSTUS_AUGUSTUS_SCRIPTS_PATH" "=" (,(string-append (assoc-ref inputs "augustus") "/bin/augustus_scripts")))
-                   `("PYTHON3_PATH" "=" (,(string-append (assoc-ref inputs "python") "/bin")))
-                   `("BAMTOOLS_PATH" "=" (,(string-append (assoc-ref inputs "bamtools") "/bin")))
-                   `("DIAMOND_PATH" "=" (,(string-append (assoc-ref inputs "diamond") "/bin")))
-                   `("BLAST_PATH" "=" (,(string-append (assoc-ref inputs "blast+") "/bin")))
-                   `("PROTHINT_PATH" "=" (,(string-append (assoc-ref inputs "prothint") "/bin")))
-                   `("SAMTOOLS_PATH" "=" (,(string-append (assoc-ref inputs "samtools") "/bin")))
-                   `("CDBTOOLS_PATH" "=" (,(string-append (assoc-ref inputs "cdbfasta") "/bin")))
-                   `("ALIGNMENT_TOOL_PATH" "=" (,(string-append (assoc-ref inputs "spaln") "/bin/spaln")
-                                                ,(string-append (assoc-ref inputs "exonerate") "/bin/exonerate")))
-                   ;`("MAKEHUB_PATH" "=" (,(string-append (assoc-ref inputs "makehub") "/bin")))
-                   )
-                 ))))
-         )))
+             (let ((bin (string-append (assoc-ref outputs "out") "/bin")))
+               (for-each
+                 (cut install-file <> bin)
+                 (find-files "scripts" "\\.(py|pl|pm)$"))
+               (for-each
+                 (cut wrap-script <>
+                      `("PERL5LIB" ":" prefix (,(getenv "PERL5LIB")
+                                               ,bin)))
+                 ;; braker.pl is the entry point so wrap it separately.
+                 (delete (string-append bin "/braker.pl")
+                         (find-files bin "\\.pl$")))
+               (wrap-script (string-append bin "/braker.pl")
+                 `("PERL5LIB" ":" prefix (,(getenv "PERL5LIB")
+                                          ,bin))
+                 `("PATH" ":" prefix (,(assoc-ref inputs "augustus")
+                                      ;,(assoc-ref inputs "genemark")
+                                      ,(assoc-ref inputs "bamtools")
+                                      ,(assoc-ref inputs "samtools")
+                                      ,(assoc-ref inputs "prothint")
+                                      ;,(assoc-ref inputs "genomethreader")
+                                      ,(assoc-ref inputs "spaln")
+                                      ,(assoc-ref inputs "exonerate")
+                                      ,(assoc-ref inputs "ncbi-blast")
+                                      ,(assoc-ref inputs "diamond")
+                                      ,(assoc-ref inputs "cdbfasta")
+                                      ,(assoc-ref inputs "gushr")
+                                      ,(assoc-ref inputs "ucsc")))
+                 ;`("GENEMARK_PATH" "=" (,(string-append (assoc-ref inputs "genemark") "/bin")))
+                 `("AUGUSTUS_BIN_PATH" "=" (,(string-append (assoc-ref inputs "augustus") "/bin")))
+                 `("AUGUSTUS_AUGUSTUS_SCRIPTS_PATH" "=" (,(string-append (assoc-ref inputs "augustus") "/share/augustus")))
+                 `("PYTHON3_PATH" "=" (,(string-append (assoc-ref inputs "python") "/bin")))
+                 `("BAMTOOLS_PATH" "=" (,(string-append (assoc-ref inputs "bamtools") "/bin")))
+                 `("DIAMOND_PATH" "=" (,(string-append (assoc-ref inputs "diamond") "/bin")))
+                 `("BLAST_PATH" "=" (,(string-append (assoc-ref inputs "blast+") "/bin")))
+                 `("PROTHINT_PATH" "=" (,(string-append (assoc-ref inputs "prothint") "/bin")))
+                 `("SAMTOOLS_PATH" "=" (,(string-append (assoc-ref inputs "samtools") "/bin")))
+                 `("CDBTOOLS_PATH" "=" (,(string-append (assoc-ref inputs "cdbfasta") "/bin")))
+                 `("ALIGNMENT_TOOL_PATH" "=" (,(string-append (assoc-ref inputs "spaln") "/bin/spaln")
+                                              ,(string-append (assoc-ref inputs "exonerate") "/bin/exonerate")))
+                 ;`("MAKEHUB_PATH" "=" (,(string-append (assoc-ref inputs "makehub") "/bin")))
+                 )))))))
     (inputs
-     `(
-       ("guile" ,guile-3.0)             ; for wrap-script
+     `(("guile" ,guile-3.0)             ; for wrap-script
 
-       ;("augustus" ,augustus)
+       ("augustus" ,augustus-bio)
        ;("genemark" ,genemark)
        ("bamtools" ,bamtools)
        ("samtools" ,samtools)
@@ -2498,10 +2495,10 @@ To run the bundled rtg-tools software you will also need java.  The
        ("exonerate" ,exonerate)
        ("ncbi-blast" ,blast+)
        ("diamond" ,diamond)
-       ("cbdfasta" ,cbdfasta)           ; provides cdbfasta and cdbyank
-       ;("gushr" ,gushr)
+       ("cdbfasta" ,cdbfasta)           ; provides cdbfasta and cdbyank
+       ("gushr" ,gushr)
        ;("makehub" ,makehub)
-       ;("ucsc" ,ucsc-genome-browser)       ; provides bin/twoBitInfo and bin/faToTwoBit
+       ("ucsc" ,ucsc-genome-browser)       ; provides bin/twoBitInfo and bin/faToTwoBit
 
        ("perl" ,perl)
        ("perl-hash-merge" ,perl-hash-merge)
@@ -2510,11 +2507,7 @@ To run the bundled rtg-tools software you will also need java.  The
        ("perl-module-load-conditional" ,perl-module-load-conditional)
        ("perl-parallel-forkmanager" ,perl-parallel-forkmanager)
        ("perl-yaml" ,perl-yaml)
-       ("python" ,python)
-       ))
-    (native-inputs
-     `(
-       ))
+       ("python" ,python)))
     (home-page "https://github.com/Gaius-Augustus/BRAKER")
     (synopsis
      "Pipeline for fully automated prediction of protein coding gene structures")
@@ -2560,7 +2553,10 @@ RNA-Seq data.")
            (delete-file "dependencies/spaln_boundary_scorer")))))
     (build-system perl-build-system)
     (arguments
-     `(#:tests? #f          ; TODO: Test suite fails, or packaging is wrong?
+     `(#:modules ((srfi srfi-26)
+                  (guix build perl-build-system)
+                  (guix build utils))
+       #:tests? #f          ; TODO: Test suite fails, or packaging is wrong?
        #:phases
        (modify-phases %standard-phases
          (add-after 'unpack 'adjust-source
@@ -2573,8 +2569,7 @@ RNA-Seq data.")
                                "/bin/spaln_boundary_scorer\""))
                (("\\$binDir/\\.\\./dependencies/spaln_table\\\"")
                 (string-append (assoc-ref inputs "spaln")
-                               "/share/spaln/table\""))
-               )))
+                               "/share/spaln/table\"")))))
          (delete 'configure)
          (delete 'build)
          (replace 'check
@@ -2589,21 +2584,20 @@ RNA-Seq data.")
            (lambda* (#:key inputs outputs #:allow-other-keys)
              (let* ((out (assoc-ref outputs "out"))
                     (bin (string-append out "/bin")))
-               (with-directory-excursion "scripts"
-                 (for-each (cut install-file <> bin)
-                           (find-files "bin" "\\.(pl|py|sh)$"))
-                 (for-each
-                   (cut wrap-script <>
-                        `("PERL5LIB" ":" prefix (,(getenv "PERL5LIB"))))
-                   (find-files bin "\\.pl$"))
-                 (for-each
-                   (cut wrap-script <>
-                        `("PATH" ":" prefix (,(assoc-ref inputs "coreutils")
-                                             ,(assoc-ref inputs "diamond")
-                                             ;,(assoc-ref inputs "genemark")
-                                             ,(assoc-ref inputs "grep")
-                                             ,(assoc-ref inputs "spaln"))))
-                   (find-files bin "\\.py$")))))))))
+               (for-each (cut install-file <> bin)
+                         (find-files "bin" "\\.(pl|py|sh)$"))
+               (for-each
+                 (cut wrap-script <>
+                      `("PERL5LIB" ":" prefix (,(getenv "PERL5LIB"))))
+                 (find-files bin "\\.pl$"))
+               (for-each
+                 (cut wrap-script <>
+                      `("PATH" ":" prefix (,(assoc-ref inputs "coreutils")
+                                           ,(assoc-ref inputs "diamond")
+                                           ;,(assoc-ref inputs "genemark")
+                                           ,(assoc-ref inputs "grep")
+                                           ,(assoc-ref inputs "spaln"))))
+                 (find-files bin "\\.py$"))))))))
     (inputs
      `(("guile" ,guile-3.0)             ; for wrap-script
 
@@ -2842,3 +2836,200 @@ are scored based on local alignment quality around their boundaries.")
     (supported-systems '("x86_64-linux"))
     ;; Licensee may use the Product solely for Licensee's own internal research purposes.
     (license license:non-copyleft)))
+
+;; Name collision with augustus in games.
+(define-public augustus-bio
+  (package
+    (name "augustus-bio")
+    (version "3.4.0")
+    (source
+      (origin
+        (method url-fetch)
+        (uri (string-append "https://github.com/Gaius-Augustus/Augustus"
+                            "/releases/download/v" version
+                            "/augustus-" version ".tar.gz"))
+        (sha256
+         (base32
+          "1j0ny6v8v3hsk76w4f4vvcqzpjjn8qfvp5q1mdanglb2g1rp4rr4"))
+        (modules '((guix build utils)))
+        (snippet
+         '(begin
+            (for-each delete-file
+                      (append
+                        '("auxprogs/aln2wig/aln2wig"
+                          "auxprogs/bam2hints/bam2hints"
+                          "auxprogs/bam2wig/bam2wig"
+                          "auxprogs/compileSpliceCands/compileSpliceCands"
+                          "auxprogs/filterBam/data/BAMseek2011July24.jar"
+                          "auxprogs/homGeneMapping/src/homGeneMapping"
+                          "auxprogs/joingenes/joingenes"
+                          "auxprogs/utrrnaseq/Debug/utrrnaseq")
+                        (find-files "." "\\.pdf$")
+                        (find-files "mansrc" "\\.1$")))
+            (for-each delete-file-recursively
+                      '("bin"
+                        "tests/__pycache__"
+                        "tests/examples/__pycache__"
+                        "tests/examples_test/__pycache__"
+                        "tests/utils/__pycache__"))))))
+    (build-system gnu-build-system)
+    (arguments
+     `(#:test-target "unit_test"
+       #:make-flags (list (string-append "CC = " ,(cc-for-target)))
+       #:phases
+       (modify-phases %standard-phases
+         (delete 'configure)        ; No configure script.
+         (add-after 'unpack 'adjust-sources
+           (lambda* (#:key inputs #:allow-other-keys)
+             (substitute* "common.mk"
+               (("AUGVERSION =.*")
+                (string-append "AUGVERSION = ",version "\n")))
+             (substitute* '("src/Makefile"
+                            "src/unittests/Makefile")
+               (("/usr/include/mysql ")
+                (string-append (assoc-ref inputs "mariadb:dev") "/include/mysql "))
+               (("/usr/include/mysql\\+\\+")
+                (string-append (assoc-ref inputs "mysql++") "/include/mysql++ "))
+               (("/usr/include/lpsolve")
+                (string-append (assoc-ref inputs "lpsolve") "/include/lpsolve")))
+             (substitute* '("auxprogs/bam2hints/Makefile"
+                            "auxprogs/filterBam/src/Makefile")
+               (("/usr/include/bamtools")
+                (string-append (assoc-ref inputs "bamtools") "/include/bamtools")))
+             (substitute* "auxprogs/bam2wig/Makefile"
+               (("/usr/include/htslib")
+                (string-append (assoc-ref inputs "htslib") "/include/htslib")))))
+         (add-after 'unpack 'remove-googletest
+           (lambda* (#:key inputs #:allow-other-keys)
+             (let ((gtest (assoc-ref inputs "googletest")))
+               (delete-file-recursively "src/googletest")
+               (substitute* "src/Makefile"
+                 (("unittest: googletest") "unittest:"))
+               (substitute* "src/unittests/Makefile"
+                 (("\\.\\./googletest/include") (string-append gtest "/include"))
+                 (("-L\\.\\./googletest") (string-append "-L" gtest "/lib"))))))
+         ;; This mimics 'make install'.
+         (replace 'install
+           (lambda* (#:key outputs #:allow-other-keys)
+             (let* ((out (assoc-ref outputs "out"))
+                    (bin (string-append out "/bin"))
+                    (scripts (string-append out "/share/augustus")))
+               (mkdir-p scripts)
+               (mkdir-p bin)
+               (copy-recursively "config" scripts)
+               (copy-recursively "bin" scripts)
+               (copy-recursively "scripts" scripts)
+               (for-each
+                 (lambda (binary)
+                   (symlink (string-append "../share/augustus/" binary)
+                            (string-append bin "/" binary)))
+                 '("augustus" "etraining" "prepareAlign" "fastBlockSearch"
+                   "load2db" "getSeq")))))
+         (add-after 'install 'install-manpages
+           (lambda* (#:key outputs #:allow-other-keys)
+             (let* ((out (assoc-ref outputs "out"))
+                    (man1 (string-append out "/share/man/man1")))
+               (for-each
+                 (lambda (adoc)
+                   (invoke "asciidoctor"
+                           "-a" "docdate=''"
+                           "-b" "manpage"
+                           (string-append "--destination-dir=" man1)
+                           adoc))
+                 (find-files "mansrc" "\\.adoc$"))))))))
+    (inputs
+     `(("bamtools" ,bamtools)
+       ("boost" ,boost)
+       ("gsl" ,gsl)
+       ("htslib" ,htslib)
+       ("lpsolve" ,lpsolve)
+       ("mariadb:lib" ,mariadb "lib")
+       ("mysql++" ,mysql++)
+       ("perl" ,perl)
+       ("python" ,python)
+       ("sqlite" ,sqlite)
+       ("suitesparse" ,suitesparse)
+       ("zlib" ,zlib)))
+    (native-inputs
+     `(("asciidoctor" ,ruby-asciidoctor)
+       ("googletest" ,googletest)
+       ("mariadb:dev" ,mariadb "dev")))
+    (home-page "http://bioinf.uni-greifswald.de/webaugustus/")
+    (synopsis "Genome annotation with AUGUSTUS")
+    (description "AUGUSTUS is a gene prediction program.  It can be used as an
+ab initio program, which means it bases its prediction purely on the sequence.
+AUGUSTUS may also incorporate hints on the gene structure coming from extrinsic
+sources such as EST, MS/MS, protein alignments and syntenic genomic alignments.")
+    (license license:artistic2.0)))
+
+;; TODO: Replace GoMoMa-1.6.2.jar with https://github.com/Jstacs/Jstacs
+(define-public gushr
+  (let ((commit "ee26d5c7eee97170c6183089d57477b338a7bc4b")
+        (revision "1"))
+    (package
+      (name "gushr")
+      (version (git-version "0.0.0" revision commit))
+      (source
+        (origin
+          (method git-fetch)
+          (uri (git-reference
+                 (url "https://github.com/Gaius-Augustus/GUSHR")
+                 (commit commit)))
+          (file-name (git-file-name name version))
+          (sha256
+           (base32
+            "00alkahjzc6zgzfm7f6zcayjvgcs3kb90fb9v855gpqcxv3y1s40"))))
+      (build-system gnu-build-system)
+      (arguments
+       `(#:phases
+         (modify-phases %standard-phases
+           (delete 'configure)          ; No configure script.
+           (delete 'build)
+           ;; Override the test.sh file.
+           (replace 'check
+             (lambda* (#:key tests? parallel-tests? inputs #:allow-other-keys)
+               (when tests?
+                 (with-directory-excursion "example"
+                   (invoke "../gushr.py"
+                           "--AUGUSTUS_SCRIPTS_PATH"
+                           (string-append (assoc-ref inputs "augustus")
+                                          "/share/augustus")
+                           "--bam" (assoc-ref inputs "RNAseq.bam")
+                           "--gtf" "augustus.gtf"
+                           "--genome" "genome.fa"
+                           "--outfile_name_stem" "gushr"
+                           "--cores" (if parallel-tests?
+                                       (number->string (parallel-job-count))
+                                       "1"))))))
+           (replace 'install
+             (lambda* (#:key inputs outputs #:allow-other-keys)
+               (let* ((out (assoc-ref outputs "out"))
+                      (bin (string-append out "/bin")))
+                 (install-file "gushr.py" bin)
+                 (install-file "GeMoMa-1.6.2.jar" bin)
+                 (wrap-script (string-append bin "/gushr.py")
+                   `("PATH" ":" prefix (,(string-append (assoc-ref inputs "java")
+                                                        "/bin")
+                                        ,(string-append (assoc-ref inputs "samtools")
+                                                        "/bin"))))))))))
+      (inputs
+       `(("guile" ,guile-3.0)      ; for wrap-script
+         ("augustus" ,augustus-bio)
+         ("java" ,icedtea-8 "jdk")
+         ("python" ,python)
+         ("samtools" ,samtools)))
+      (native-inputs
+       `(("RNAseq.bam"
+          ,(origin
+             (method url-fetch)
+             (uri "http://bioinf.uni-greifswald.de/bioinf/braker/RNAseq.bam")
+             (sha256
+              (base32
+               "1apmgh9irwhfnbpqwvcjasfs524g03i284bxbb8czdkqfsa74w47"))))))
+      (home-page "https://github.com/Gaius-Augustus/GUSHR")
+      (synopsis "Generating UTRs from SHort Reads")
+      (description "Assembly-free construction of UTRs from short read RNA-Seq
+data on the basis of coding sequence annotation.  This tool has been adapted to
+the format needs of AUGUSTUS/BRAKER and employs GeMoMa for generating UTRs from
+RNA-Seq coverage data.")
+      (license license:gpl3+))))
diff --git a/gn/packages/databases.scm b/gn/packages/databases.scm
index 93280c2..9c4a109 100644
--- a/gn/packages/databases.scm
+++ b/gn/packages/databases.scm
@@ -2,6 +2,7 @@
   #:use-module (gnu packages)
   #:use-module (guix packages)
   #:use-module (guix download)
+  #:use-module ((guix licenses) #:prefix license:)
   #:use-module (guix utils)
   #:use-module (guix build-system gnu)
   #:use-module (guix build-system python)
@@ -136,3 +137,57 @@
       (inputs
        `(("zlib" ,zlib)
          ,@(package-inputs base))))))
+
+(define-public mysql++
+  (package
+    (name "mysql++")
+    (version "3.3.0")
+    (source
+      (origin
+        (method url-fetch)
+        (uri (string-append "https://tangentsoft.com/mysqlpp/releases"
+                            "/mysql++-" version ".tar.gz"))
+        (sha256
+         (base32
+          "1kz7l1ngk649cpp2h1cnyqan9px8d50r0dk7kngwrhkcam3br724"))))
+    (build-system gnu-build-system)
+    (arguments
+     `(#:configure-flags
+       (list (string-append "--with-mysql-include="
+                            (assoc-ref %build-inputs "mariadb:dev"))
+             (string-append "--with-mysql-lib="
+                            (assoc-ref %build-inputs "mariadb:lib")))
+       #:phases
+       (modify-phases %standard-phases
+         ;; It is unclear how to run the test suite so we just invoke the
+         ;; compiled binaries which start with 'test_*'.
+         (replace 'check
+           (lambda* (#:key tests? #:allow-other-keys)
+             (when tests?
+               (setenv "LD_PRELOAD" "./libmysqlpp.so.3")
+               (and
+                 (invoke "./test_array_index")
+                 (invoke "./test_cpool")
+                 (invoke "./test_datetime")
+                 (invoke "./test_insertpolicy")
+                 (invoke "./test_inttypes")
+                 (invoke "./test_manip")
+                 (invoke "./test_null_comparison")
+                 (invoke "./test_qssqls")
+                 (invoke "./test_qstream")
+                 (invoke "./test_query_copy")
+                 (invoke "./test_sqlstream")
+                 (invoke "./test_ssqls2")
+                 (invoke "./test_string")
+                 ;(invoke "./test_tcp")     ; Requires TCP connection.
+                 (invoke "./test_uds")
+                 (invoke "./test_wnp")))
+             #t)))))
+    (inputs
+     `(("mariadb:dev" ,mariadb "dev")
+       ("mariadb:lib" ,mariadb "lib")))
+    (home-page "https://tangentsoft.net/mysqlpp/")
+    (synopsis "MySQL C++ library bindings")
+    (description "MySQL++ is a complex C++ API for MySQL.  The goal of this API
+is to make working with Queries as easy as working with other STL containers.")
+    (license license:lgpl2.1+)))