Skip to content

Commit

Permalink
feat: hardcode organism assembly count as 1 (#177)
Browse files Browse the repository at this point in the history
  • Loading branch information
@hunterckx
hunterckx committed Dec 12, 2024
1 parent 4f9dfd6 commit d326bb3
Show file tree
Hide file tree
Showing 3 changed files with 41 additions and 41 deletions.
2 changes: 1 addition & 1 deletion files/build-files-from-ncbi.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@ def get_organism_row(organism_info, accession):
return {
"taxon": organism_taxonomy["current_scientific_name"]["name"],
"taxonomyId": str(organism_taxonomy["tax_id"]),
"assemblyCount": next(count["count"] for count in organism_taxonomy["counts"] if count["type"] == "COUNT_TYPE_ASSEMBLY"),
"assemblyCount": 1,
"accession": accession,
}

Expand Down
40 changes: 20 additions & 20 deletions files/out/organisms.json
View file
Original file line number Diff line number Diff line change
@@ -1,30 +1,30 @@
[
{
"assemblyCount": 6911,
"assemblyCount": 1,
"ncbiTaxonomyId": "10244",
"tags": [
"Virus"
],
"taxon": "Monkeypox virus"
},
{
"assemblyCount": 7829,
"assemblyCount": 1,
"ncbiTaxonomyId": "1773",
"tags": [
"Bact"
],
"taxon": "Mycobacterium tuberculosis"
},
{
"assemblyCount": 13,
"assemblyCount": 1,
"ncbiTaxonomyId": "199306",
"tags": [
"VEuPathDB"
],
"taxon": "Coccidioides posadasii"
},
{
"assemblyCount": 12408,
"assemblyCount": 1,
"ncbiTaxonomyId": "2697049",
"tags": [
"Virus"
Expand All @@ -38,121 +38,121 @@
"taxon": "Culex pipiens pallens"
},
{
"assemblyCount": 183,
"assemblyCount": 1,
"ncbiTaxonomyId": "5207",
"tags": [],
"taxon": "Cryptococcus neoformans"
},
{
"assemblyCount": 117,
"assemblyCount": 1,
"ncbiTaxonomyId": "5476",
"tags": [],
"taxon": "Candida albicans"
},
{
"assemblyCount": 5,
"assemblyCount": 1,
"ncbiTaxonomyId": "5501",
"tags": [],
"taxon": "Coccidioides immitis"
},
{
"assemblyCount": 11,
"assemblyCount": 1,
"ncbiTaxonomyId": "5660",
"tags": [
"VEuPathDB"
],
"taxon": "Leishmania braziliensis"
},
{
"assemblyCount": 12,
"assemblyCount": 1,
"ncbiTaxonomyId": "5661",
"tags": [
"VEuPathDB"
],
"taxon": "Leishmania donovani"
},
{
"assemblyCount": 7,
"assemblyCount": 1,
"ncbiTaxonomyId": "5664",
"tags": [
"VEuPathDB"
],
"taxon": "Leishmania major"
},
{
"assemblyCount": 6,
"assemblyCount": 1,
"ncbiTaxonomyId": "5691",
"tags": [
"VEuPathDB"
],
"taxon": "Trypanosoma brucei"
},
{
"assemblyCount": 44,
"assemblyCount": 1,
"ncbiTaxonomyId": "5693",
"tags": [
"VEuPathDB"
],
"taxon": "Trypanosoma cruzi"
},
{
"assemblyCount": 29,
"assemblyCount": 1,
"ncbiTaxonomyId": "5811",
"tags": [
"VEuPathDB"
],
"taxon": "Toxoplasma gondii"
},
{
"assemblyCount": 67,
"assemblyCount": 1,
"ncbiTaxonomyId": "5833",
"tags": [
"VEuPathDB"
],
"taxon": "Plasmodium falciparum"
},
{
"assemblyCount": 19,
"assemblyCount": 1,
"ncbiTaxonomyId": "5855",
"tags": [
"VEuPathDB"
],
"taxon": "Plasmodium vivax"
},
{
"assemblyCount": 10,
"assemblyCount": 1,
"ncbiTaxonomyId": "5860",
"tags": [
"VEuPathDB"
],
"taxon": "Plasmodium vinckei"
},
{
"assemblyCount": 15,
"assemblyCount": 1,
"ncbiTaxonomyId": "5861",
"tags": [
"VEuPathDB"
],
"taxon": "Plasmodium yoelii"
},
{
"assemblyCount": 7,
"assemblyCount": 1,
"ncbiTaxonomyId": "7165",
"tags": [
"VEuPathDB"
],
"taxon": "Anopheles gambiae"
},
{
"assemblyCount": 3,
"assemblyCount": 1,
"ncbiTaxonomyId": "7176",
"tags": [
"VEuPathDB"
],
"taxon": "Culex quinquefasciatus"
},
{
"assemblyCount": 352,
"assemblyCount": 1,
"ncbiTaxonomyId": "746128",
"tags": [],
"taxon": "Aspergillus fumigatus"
Expand Down
40 changes: 20 additions & 20 deletions files/source/organisms-from-ncbi.tsv
View file
Original file line number Diff line number Diff line change
@@ -1,22 +1,22 @@
taxon taxonomyId assemblyCount accession CustomTags
Plasmodium falciparum 5833 67 GCF_000002765.6 VEuPathDB
Plasmodium vivax 5855 19 GCF_000002415.2 VEuPathDB
Plasmodium yoelii 5861 15 GCF_900002385.2 VEuPathDB
Plasmodium vinckei 5860 10 GCF_900681995.1 VEuPathDB
Plasmodium falciparum 5833 1 GCF_000002765.6 VEuPathDB
Plasmodium vivax 5855 1 GCF_000002415.2 VEuPathDB
Plasmodium yoelii 5861 1 GCF_900002385.2 VEuPathDB
Plasmodium vinckei 5860 1 GCF_900681995.1 VEuPathDB
Culex pipiens pallens 42434 1 GCF_016801865.2
Culex quinquefasciatus 7176 3 GCF_015732765.1 VEuPathDB
Anopheles gambiae 7165 7 GCF_943734735.2 VEuPathDB
Toxoplasma gondii 5811 29 GCF_000006565.2 VEuPathDB
Mycobacterium tuberculosis 1773 7829 GCF_000195955.2 Bact
Coccidioides posadasii 199306 13 GCF_018416015.2 VEuPathDB
Coccidioides immitis 5501 5 GCF_000149335.2
Trypanosoma cruzi 5693 44 GCF_000209065.1 VEuPathDB
Trypanosoma brucei 5691 6 GCF_000002445.2 VEuPathDB
Leishmania major 5664 7 GCF_000002725.2 VEuPathDB
Leishmania donovani 5661 12 GCF_000227135.1 VEuPathDB
Leishmania braziliensis 5660 11 GCF_000002845.2 VEuPathDB
Severe acute respiratory syndrome coronavirus 2 2697049 12408 GCF_009858895.2 Virus
Monkeypox virus 10244 6911 GCF_000857045.1 Virus
Aspergillus fumigatus 746128 352 GCF_000002655.1
Candida albicans 5476 117 GCF_000182965.3
Cryptococcus neoformans 5207 183 GCF_000091045.1
Culex quinquefasciatus 7176 1 GCF_015732765.1 VEuPathDB
Anopheles gambiae 7165 1 GCF_943734735.2 VEuPathDB
Toxoplasma gondii 5811 1 GCF_000006565.2 VEuPathDB
Mycobacterium tuberculosis 1773 1 GCF_000195955.2 Bact
Coccidioides posadasii 199306 1 GCF_018416015.2 VEuPathDB
Coccidioides immitis 5501 1 GCF_000149335.2
Trypanosoma cruzi 5693 1 GCF_000209065.1 VEuPathDB
Trypanosoma brucei 5691 1 GCF_000002445.2 VEuPathDB
Leishmania major 5664 1 GCF_000002725.2 VEuPathDB
Leishmania donovani 5661 1 GCF_000227135.1 VEuPathDB
Leishmania braziliensis 5660 1 GCF_000002845.2 VEuPathDB
Severe acute respiratory syndrome coronavirus 2 2697049 1 GCF_009858895.2 Virus
Monkeypox virus 10244 1 GCF_000857045.1 Virus
Aspergillus fumigatus 746128 1 GCF_000002655.1
Candida albicans 5476 1 GCF_000182965.3
Cryptococcus neoformans 5207 1 GCF_000091045.1

0 comments on commit d326bb3

Please sign in to comment.