From 83985b3a5beb9060332b70796757ea62a0c59993 Mon Sep 17 00:00:00 2001
From: willtyler <will.tyler11@gmail.com>
Date: Fri, 30 Aug 2024 19:53:03 +0000
Subject: [PATCH] Add provenance header

---
 tests/test_cli.py           | 13 +++++++++++++
 tests/test_vcf_roundtrip.py |  2 +-
 tests/test_vcf_writer.py    |  8 ++++----
 vcztools/cli.py             |  1 +
 vcztools/vcf_writer.py      | 15 +++++++++++++--
 5 files changed, 32 insertions(+), 7 deletions(-)
 create mode 100644 tests/test_cli.py

diff --git a/tests/test_cli.py b/tests/test_cli.py
new file mode 100644
index 0000000..c5387a2
--- /dev/null
+++ b/tests/test_cli.py
@@ -0,0 +1,13 @@
+import pathlib
+import re
+
+from tests.test_bcftools_validation import run_vcztools
+from tests.utils import vcz_path_cache
+
+
+def test_version_header():
+    vcf_path = pathlib.Path("tests/data/vcf/sample.vcf.gz")
+    vcz_path = vcz_path_cache(vcf_path)
+
+    output = run_vcztools(f"view {vcz_path}")
+    assert re.search("##vcztools_viewCommand=.+; Date=.+", output)
diff --git a/tests/test_vcf_roundtrip.py b/tests/test_vcf_roundtrip.py
index 329c161..b11593a 100644
--- a/tests/test_vcf_roundtrip.py
+++ b/tests/test_vcf_roundtrip.py
@@ -21,5 +21,5 @@ def test_vcf_to_zarr_to_vcf__real_files(tmp_path, vcf_file):
     original = pathlib.Path("tests/data/vcf") / vcf_file
     vcz = vcz_path_cache(original)
     generated = tmp_path.joinpath("output.vcf")
-    write_vcf(vcz, generated)
+    write_vcf(vcz, generated, no_version=True)
     assert_vcfs_close(original, generated)
diff --git a/tests/test_vcf_writer.py b/tests/test_vcf_writer.py
index d1f79ed..f493db3 100644
--- a/tests/test_vcf_writer.py
+++ b/tests/test_vcf_writer.py
@@ -18,10 +18,10 @@ def test_write_vcf(tmp_path, output_is_path):
     output = tmp_path.joinpath("output.vcf")
 
     if output_is_path:
-        write_vcf(vcz, output)
+        write_vcf(vcz, output, no_version=True)
     else:
         output_str = StringIO()
-        write_vcf(vcz, output_str)
+        write_vcf(vcz, output_str, no_version=True)
         with open(output, "w") as f:
             f.write(output_str.getvalue())
 
@@ -251,10 +251,10 @@ def test_write_vcf__header_flags(tmp_path):
     output = tmp_path.joinpath("output.vcf")
 
     output_header = StringIO()
-    write_vcf(vcz, output_header, header_only=True)
+    write_vcf(vcz, output_header, header_only=True, no_version=True)
 
     output_no_header = StringIO()
-    write_vcf(vcz, output_no_header, no_header=True)
+    write_vcf(vcz, output_no_header, no_header=True, no_version=True)
     assert not output_no_header.getvalue().startswith("#")
 
     # combine outputs and check VCFs match
diff --git a/vcztools/cli.py b/vcztools/cli.py
index 72de451..10f9c89 100644
--- a/vcztools/cli.py
+++ b/vcztools/cli.py
@@ -130,6 +130,7 @@ def view(
         output,
         header_only=header_only,
         no_header=no_header,
+        no_version=no_version,
         variant_regions=regions,
         variant_targets=targets,
         samples=samples,
diff --git a/vcztools/vcf_writer.py b/vcztools/vcf_writer.py
index 871b536..51e7d8a 100644
--- a/vcztools/vcf_writer.py
+++ b/vcztools/vcf_writer.py
@@ -1,6 +1,8 @@
 import functools
 import io
 import re
+import sys
+from datetime import datetime
 from typing import Optional
 
 import numpy as np
@@ -82,6 +84,7 @@ def write_vcf(
     vcf_header: Optional[str] = None,
     header_only: bool = False,
     no_header: bool = False,
+    no_version: bool = False,
     variant_regions=None,
     variant_targets=None,
     samples=None,
@@ -157,7 +160,9 @@ def write_vcf(
                 original_header = root.attrs["vcf_header"]
             else:
                 original_header = None
-            vcf_header = _generate_header(root, original_header, sample_ids)
+            vcf_header = _generate_header(
+                root, original_header, sample_ids, no_version=no_version
+            )
 
         if not no_header:
             print(vcf_header, end="", file=output)
@@ -382,7 +387,7 @@ def c_chunk_to_vcf(
         print(line, file=output)
 
 
-def _generate_header(ds, original_header, sample_ids):
+def _generate_header(ds, original_header, sample_ids, *, no_version: bool = False):
     output = io.StringIO()
 
     contigs = list(ds["contig_id"][:])
@@ -509,6 +514,12 @@ def _generate_header(ds, original_header, sample_ids):
         else:
             print(f"##contig=<ID={contig},length={contig_lengths[i]}>", file=output)
 
+    if not no_version:
+        print(
+            f"##vcztools_viewCommand={' '.join(sys.argv[1:])}; Date={datetime.now()}",
+            file=output,
+        )
+
     # [1.5 Header line syntax]
     print(
         "#CHROM",