From 77f3ab4586ded22fbda055bea3bce7f8801d09cc Mon Sep 17 00:00:00 2001
From: Dave Lawrence <davmlaw@gmail.com>
Date: Mon, 6 Mar 2023 13:27:50 +1030
Subject: [PATCH] Move GFF out of cdot package (only used to build data). #35 -
 make tags genome build specific

---
 cdot/__init__.py                                     | 2 +-
 {cdot/gff => generate_transcript_data}/__init__.py   | 0
 generate_transcript_data/cdot_json.py                | 2 +-
 {cdot/gff => generate_transcript_data}/gff_parser.py | 2 +-
 setup.cfg                                            | 5 ++++-
 tests/test_gff_parsers.py                            | 2 +-
 6 files changed, 8 insertions(+), 5 deletions(-)
 rename {cdot/gff => generate_transcript_data}/__init__.py (100%)
 rename {cdot/gff => generate_transcript_data}/gff_parser.py (99%)

diff --git a/cdot/__init__.py b/cdot/__init__.py
index 0131ebc..7130536 100644
--- a/cdot/__init__.py
+++ b/cdot/__init__.py
@@ -1,4 +1,4 @@
-__version__ = "0.2.13"
+__version__ = "0.2.14"
 
 
 def get_json_schema_version():
diff --git a/cdot/gff/__init__.py b/generate_transcript_data/__init__.py
similarity index 100%
rename from cdot/gff/__init__.py
rename to generate_transcript_data/__init__.py
diff --git a/generate_transcript_data/cdot_json.py b/generate_transcript_data/cdot_json.py
index 7674b0a..0cd262c 100755
--- a/generate_transcript_data/cdot_json.py
+++ b/generate_transcript_data/cdot_json.py
@@ -13,7 +13,7 @@
 import cdot
 import ijson
 from cdot.pyhgvs.pyhgvs_transcript import PyHGVSTranscriptFactory
-from cdot.gff.gff_parser import GTFParser, GFF3Parser
+from gff_parser import GTFParser, GFF3Parser
 from cdot.json_encoders import SortedSetEncoder
 from pyhgvs import CDNACoord
 
diff --git a/cdot/gff/gff_parser.py b/generate_transcript_data/gff_parser.py
similarity index 99%
rename from cdot/gff/gff_parser.py
rename to generate_transcript_data/gff_parser.py
index 82efea7..704dd43 100644
--- a/cdot/gff/gff_parser.py
+++ b/generate_transcript_data/gff_parser.py
@@ -66,7 +66,7 @@ def _finish(self):
             gene_data["url"] = self.url
 
         # At the moment the transcript dict is flat - need to move it into "genome_builds" dict
-        GENOME_BUILD_FIELDS = ["cds_start", "cds_end", "strand", "contig", "exons", "other_chroms"]
+        GENOME_BUILD_FIELDS = ["cds_start", "cds_end", "strand", "contig", "exons", "other_chroms", "tag"]
         for transcript_accession, transcript_data in self.transcript_data_by_accession.items():
             if protein := self.transcript_proteins.get(transcript_accession):
                 transcript_data["protein"] = protein
diff --git a/setup.cfg b/setup.cfg
index a1be9ec..22c4142 100644
--- a/setup.cfg
+++ b/setup.cfg
@@ -26,4 +26,7 @@ install_requires =
 
 [options.packages.find]
 where =
-exclude=tests
+exclude=
+    tests
+    generate_transcript_data
+
diff --git a/tests/test_gff_parsers.py b/tests/test_gff_parsers.py
index d44d381..77dce04 100644
--- a/tests/test_gff_parsers.py
+++ b/tests/test_gff_parsers.py
@@ -2,7 +2,7 @@
 import os
 from inspect import getsourcefile
 import unittest
-from cdot.gff.gff_parser import GTFParser, GFF3Parser
+from generate_transcript_data.gff_parser import GTFParser, GFF3Parser
 
 
 class Test(unittest.TestCase):