Added e2e test support, minor code refactoring

examples/input_example.tsv

@@ -1,3 +1,3 @@
-genome_name	genome_path	accessions	assembly_software	binning_software	binning_parameters	stats_generation_software	completeness	contamination	genome_coverage	metagenome	co-assembly	broad_environment	local_environment	environmental_medium	rRNA_presence	taxonomy_lineage
+genome_name	genome_path	accessions	assembly_software	binning_software	binning_parameters	stats_generation_software	completeness	contamination	genome_coverage	metagenome	co-assembly	broad_environment	local_environment	environmental_medium	rRNA_presence	NCBI_lineage
 coassembly2728_3.fa	./coassembly2728_3.fa.gz	SRR1622827,SRR1622828	MetaSPAdes_v3.15.3	metaWRAP_v1.3	default	EukCC_v2.1.0	98.2	0.2	26.24	skin metagenome	True	human	skin	finger	False	1;131567;2759;33154;4751;451864;5204;452284;1538075;162474;742845;55193;76775
 ERR4647712_megahit_bin.fa	./ERR4647712_megahit_bin.fa.gz	ERR4647712	megahit_v1.2.9	MGnify-genomes-generation-pipeline_v1.0.0	default	CheckM2_v1.0.1	90.81	0.59	14.2	chicken gut metagenome	False	chicken	gut	mucosa	True	d__Bacteria;p__Firmicutes;c__Bacilli;o__Lactobacillales;f__Lactobacillaceae;g__Lactobacillus;s__Lactobacillus crispatus

genomeuploader/genome_upload.py

@@ -173,11 +173,8 @@ def round_stats(stats):
 def compute_MAG_quality(completeness, contamination, RNApresence):
     RNApresent = str(RNApresence).lower() in ["true", "yes", "y"]
     quality = MQ
-    if completeness >= 90 and contamination <= 5 and RNApresent:
+    if float(completeness) >= 90 and float(contamination) <= 5 and RNApresent:
         quality = HQ
-
-    completeness = str(round_stats(completeness))
-    contamination = str(round_stats(contamination))
 
     return quality, completeness, contamination
 
@@ -333,12 +330,12 @@ def extract_genomes_info(inputFile, genomeType, live):
         genomeInfo[gen]["isolationSource"] = genomeInfo[gen]["metagenome"]
 
         try:
-            genomeInfo[gen]["completeness"] = str(round_stats(genomeInfo[gen]["completeness"]))
-            genomeInfo[gen]["contamination"] = str(round_stats(genomeInfo[gen]["contamination"]))
-
-            quality = compute_MAG_quality(genomeInfo[gen]["completeness"],
-                 genomeInfo[gen]["contamination"], genomeInfo[gen]["rRNA_presence"])
-            genomeInfo[gen]["MAG_quality"] = quality
+            (genomeInfo[gen]["MAG_quality"], 
+            genomeInfo[gen]["completeness"], 
+            genomeInfo[gen]["contamination"]) = compute_MAG_quality(
+                                    str(round_stats(genomeInfo[gen]["completeness"])),
+                                    str(round_stats(genomeInfo[gen]["contamination"])), 
+                                    genomeInfo[gen]["rRNA_presence"])
         except IndexError:
             pass
 

pyproject.toml

@@ -26,6 +26,14 @@ dependencies = [
     "python-dotenv==1.0.1"
 ]
 
+[tool.pytest.ini_options]
+minversion = "6.0"
+addopts = "-ra -q"
+testpaths = [
+    "tests",
+    "integration",
+]
+
 [project.optional-dependencies]
 dev = [
     "pre-commit==3.3.3",
@@ -59,9 +67,6 @@ ignore = [
     "S308", # flake8-bandit suspicious-mark-safe-usage
     "E501", # pycodestyle line-too-long
 ]
-line-length = 140
-src = ["fetchtool", "tests"]
-target-version = "py38"
 
 [tool.ruff.flake8-pytest-style]
 fixture-parentheses = false
@@ -83,5 +88,5 @@ current_version = "0.9.0"
 commit = true
 tag = true
 
-[[tool.bumpversion.files]]
+[tool.bumpversion.files]
 filename = "genomeuploader/__init__.py"

requirements-test.yml

@@ -0,0 +1 @@
+-r requirements.yml

tests/fixtures/ERR6769700_bin.1.fa.gz

@@ -0,0 +1,4 @@
+>contig1
+AGAGTAGGCGAGCTAGCGAAC
+>contig2
+GATGAGCTACGAGCGATTAAC

tests/fixtures/bin_upload/bin_backup.json

@@ -0,0 +1,12 @@
+{
+    "ERR6769700": {
+        "instrumentModel": "DNBSEQ-G400", 
+        "collectionDate": "2019-10-15", 
+        "country": "Spain", 
+        "latitude": "41.170048", 
+        "longitude": "1.168491", 
+        "projectDescription": "Metagenomic raw reads, assemblies, and bins derived from HoloFood chicken ileum samples. Samples selected for this batch were deeply sequenced, and were randomised among trials (feed), age, and breed to overcome batch effect. The samples in this project contributed to the chicken ileum MAG catalogue (project: PRJEB55375 [ERP140264]).", 
+        "study": "ERP131894", 
+        "assembler": "not provided", 
+        "sampleAccession": "SAMEA10130103"}
+}

tests/fixtures/input_fixture.tsv

@@ -0,0 +1,2 @@
+genome_name	genome_path	accessions	assembly_software	binning_software	binning_parameters	stats_generation_software	completeness	contamination	genome_coverage	metagenome	co-assembly	broad_environment	local_environment	environmental_medium	rRNA_presence	NCBI_lineage
+ERR6769700_bin.1	./tests/fixtures/ERR6769700_bin.1.fa.gz	ERR6769700	megahit_v1.2.9	MGnify-genomes-generation-pipeline_v1.0.0	default	CheckM2_v1.0.1	90.81314	0.59	14.2	chicken gut metagenome	False	chicken	gut	mucosa	True	d__Bacteria;p__Firmicutes;c__Bacilli;o__Lactobacillales;f__Lactobacillaceae;g__Lactobacillus;s__Lactobacillus crispatus

tests/test_checkOutcome_e2e.yaml

@@ -0,0 +1,7 @@
+- name: genomeuploader end to end test
+  command: "python genomeuploader/genome_upload.py -u ERP159782 --genome_info tests/fixtures/input_fixture.tsv --out tests/fixtures/ --bins --centre_name EMG"
+  files:
+    - path: "tests/fixtures/bin_upload/manifests/"
+    - path: "tests/fixtures/bin_upload/genome_samples.xml"
+    - path: "tests/fixtures/bin_upload/registered_bins_test.tsv"
+    - path: "tests/fixtures/bin_upload/submission.xml"