From 2c621522b58a72e95b3953d7633211c7e14b5cfe Mon Sep 17 00:00:00 2001
From: Kenzo-Hugo Hillion <kenzo-hugo.hillion1@pasteur.fr>
Date: Mon, 10 Feb 2020 16:15:58 +0100
Subject: [PATCH] start Snakefile to split fasta
---
tools/utils/split_fasta/Snakefile | 20 ++++++++++++++++++++
tools/utils/split_fasta/config_example.yaml | 0
2 files changed, 20 insertions(+)
create mode 100644 tools/utils/split_fasta/Snakefile
create mode 100644 tools/utils/split_fasta/config_example.yaml
diff --git a/tools/utils/split_fasta/Snakefile b/tools/utils/split_fasta/Snakefile
new file mode 100644
index 0000000..d4bd950
--- /dev/null
+++ b/tools/utils/split_fasta/Snakefile
@@ -0,0 +1,20 @@
+__split_fasta_number_sequences = config.get('split_fasta', {}).get('number_sequences', 1000000)
+__split_fasta_prefix = config.get('split_fasta', {}).get('prefix', 'seq_chunk_')
+
+EXPECTED_EXT = [f"{i:05d}" for i in range(0, int(9898412/__split_fasta_number_sequences) + 1)]
+
+rule split_fasta:
+ """
+ Split a FASTA file with the desired number of sequences per chunk
+ """
+ input:
+ __split_fasta_input
+ output:
+ __split_fasta_output
+ params:
+ n_lines = __split_fasta_number_sequences * 2,
+ prefix = __split_fasta_prefix
+ shell:
+ """
+ cat {input} | awk '/^>/ {{if(N>0) printf("\\n"); printf("%s\\n",$0);++N;next;}} {{ printf("%s",$0);}} END {{printf("\\n");}}' | split -l {params.n_lines} -a 5 -d - {params.prefix}
+ """
diff --git a/tools/utils/split_fasta/config_example.yaml b/tools/utils/split_fasta/config_example.yaml
new file mode 100644
index 0000000..e69de29
--
GitLab