Merge branch 'latest' into add/cli_parse_args

sourmash-bio · Aug 12, 2023 · e9c0f8c · e9c0f8c
2 parents bbf3d64 + f5c72fc
commit e9c0f8c
Show file tree

Hide file tree

Showing 8 changed files with 53 additions and 13 deletions.
diff --git a/.github/workflows/python.yml b/.github/workflows/python.yml
@@ -57,7 +57,7 @@ jobs:
 
       - name: Start Redis
         if: startsWith(runner.os, 'Linux') && (matrix.py == '3.9')
-        uses: supercharge/redis-github-action@1.5.0
+        uses: supercharge/redis-github-action@1.6.0
         with:
           redis-version: 6
 

diff --git a/Cargo.lock b/Cargo.lock
diff --git a/doc/command-line.md b/doc/command-line.md
@@ -612,7 +612,7 @@ sourmash tax metagenome
     --gather-csv HSMA33MX_gather_x_gtdbrs202_k31.csv \
     --gather-csv PSM6XBW3_gather_x_gtdbrs202_k31.csv \
     --taxonomy gtdb-rs202.taxonomy.v2.csv \
-    --output-format krona --rank species
+    --output-format lineage_summary --rank species
 ```
 
 example `lineage_summary`:

diff --git a/doc/databases.md b/doc/databases.md
@@ -42,9 +42,9 @@ The GTDB genomic representatives are a low-redundancy subset of Genbank genomes,
 
 | K-mer size | Zipfile collection | SBT | LCA |
 | -------- | -------- | -------- | ---- |
-| 21 | [download (2.2 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k21.zip) | [download (4.4 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k21.sbt.zip) | [download (189 MB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs207gtdb-rs214-reps.k21.lca.json.gz) |
-| 31 | [download (2.2 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k31.zip) | [download (4.4 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k31.sbt.zip) | [download (221 MB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs207/gtdb-rs214-reps.k31.lca.json.gz) |
-| 51 | [download (2.2 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k51.zip) | [download (4.4 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k51.sbt.zip) | [download (230 MB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs207/gtdb-rs214-reps.k51.lca.json.gz) |
+| 21 | [download (2.2 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k21.zip) | [download (4.4 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k21.sbt.zip) | [download (189 MB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k21.lca.json.gz) |
+| 31 | [download (2.2 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k31.zip) | [download (4.4 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k31.sbt.zip) | [download (221 MB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k31.lca.json.gz) |
+| 51 | [download (2.2 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k51.zip) | [download (4.4 GB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k51.sbt.zip) | [download (230 MB)](https://farm.cse.ucdavis.edu/~ctbrown/sourmash-db/gtdb-rs214/gtdb-rs214-reps.k51.lca.json.gz) |
 
 ### GTDB R08-RS214 all genomes (403k)
 

diff --git a/src/core/Cargo.toml b/src/core/Cargo.toml
@@ -42,7 +42,7 @@ num-iter = "0.1.43"
 once_cell = "1.18.0" # once_cell 1.14+ requires Rust 1.56+
 rayon = { version = "1.7.0", optional = true }
 serde = { version = "1.0.168", features = ["derive"] }
-serde_json = "1.0.103"
+serde_json = "1.0.104"
 primal-check = "0.3.1"
 thiserror = "1.0"
 typed-builder = "0.14.0"
@@ -58,7 +58,7 @@ criterion = "0.5.1"
 needletail = { version = "0.5.1", default-features = false }
 proptest = { version = "1.2.0", default-features = false, features = ["std"]}
 rand = "0.8.2"
-tempfile = "3.7.0"
+tempfile = "3.7.1"
 
 [[bench]]
 name = "index"

diff --git a/src/sourmash/cli/sig/split.py b/src/sourmash/cli/sig/split.py
@@ -59,6 +59,10 @@ def subparser(subparsers):
         '--from-file',
         help='a text file containing a list of files to load signatures from'
     )
+    subparser.add_argument(
+        '-E', '--extension', type=str, default='.sig',
+        help="write files with this extension ('.sig' by default)"
+    )
     add_ksize_arg(subparser)
     add_moltype_args(subparser)
     add_picklist_args(subparser)

diff --git a/src/sourmash/sig/__main__.py b/src/sourmash/sig/__main__.py
@@ -139,8 +139,8 @@ def split(args):
     _extend_signatures_with_from_file(args)
 
     output_names = set()
-    output_scaled_template = '{md5sum}.k={ksize}.scaled={scaled}.{moltype}.dup={dup}.{basename}.sig'
-    output_num_template = '{md5sum}.k={ksize}.num={num}.{moltype}.dup={dup}.{basename}.sig'
+    output_scaled_template = '{md5sum}.k={ksize}.scaled={scaled}.{moltype}.dup={dup}.{basename}' + args.extension
+    output_num_template = '{md5sum}.k={ksize}.num={num}.{moltype}.dup={dup}.{basename}' + args.extension
 
     if args.output_dir:
         if not os.path.exists(args.output_dir):

diff --git a/tests/test_cmd_signature.py b/tests/test_cmd_signature.py
@@ -1465,6 +1465,42 @@ def test_sig_split_3_multisig(c):
         assert os.path.exists(c.output(filename))
 
 
+def test_sig_split_3_multisig_sig_gz(runtmp):
+    # split 47 and 47+63-multisig.sig with a .sig.gz extension
+    c = runtmp
+
+    sig47 = utils.get_test_data('47.fa.sig')
+    multisig = utils.get_test_data('47+63-multisig.sig')
+    c.run_sourmash('sig', 'split', sig47, multisig, '-E', '.sig.gz')
+
+    outlist = ['57e2b22f.k=31.scaled=1000.DNA.dup=0.none.sig.gz',
+               'bde81a41.k=31.scaled=1000.DNA.dup=0.none.sig.gz',
+               'f033bbd8.k=31.scaled=1000.DNA.dup=0.none.sig.gz',
+               '87a9aec4.k=31.scaled=1000.DNA.dup=0.none.sig.gz',
+               '837bf2a7.k=31.scaled=1000.DNA.dup=0.none.sig.gz',
+               '485c3377.k=31.scaled=1000.DNA.dup=0.none.sig.gz']
+    for filename in outlist:
+        assert os.path.exists(c.output(filename))
+
+
+def test_sig_split_3_multisig_zip(runtmp):
+    # split 47 and 47+63-multisig.sig with a .zip extension
+    c = runtmp
+
+    sig47 = utils.get_test_data('47.fa.sig')
+    multisig = utils.get_test_data('47+63-multisig.sig')
+    c.run_sourmash('sig', 'split', sig47, multisig, '-E', '.zip')
+
+    outlist = ['57e2b22f.k=31.scaled=1000.DNA.dup=0.none.zip',
+               'bde81a41.k=31.scaled=1000.DNA.dup=0.none.zip',
+               'f033bbd8.k=31.scaled=1000.DNA.dup=0.none.zip',
+               '87a9aec4.k=31.scaled=1000.DNA.dup=0.none.zip',
+               '837bf2a7.k=31.scaled=1000.DNA.dup=0.none.zip',
+               '485c3377.k=31.scaled=1000.DNA.dup=0.none.zip']
+    for filename in outlist:
+        assert os.path.exists(c.output(filename))
+
+
 @utils.in_tempdir
 def test_sig_split_4_sbt_prot(c):
     # split sbt