split iso2 output for regex - fix 113

IndEcol · Aug 5, 2022 · 213fd8a · 213fd8a
1 parent 03cc3a0
commit 213fd8a
Show file tree

Hide file tree

Showing 5 changed files with 38 additions and 4 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,14 @@
 # Change Log
 
+
+## 0.7.7 - 20220805
+
+
+### Bug fixes
+
+- ISO2 did return the regex for GB|UK - fix #113
+
+
 ## 0.7.6 - 20220802
 
 ### Breaking

diff --git a/country_converter/country_converter.py b/country_converter/country_converter.py
@@ -470,7 +470,14 @@ def data_loader(data):
         # class.
         def fun_provider(df, datacol):
             def fun_provided(to):
-                return df.loc[:, [to, datacol]].dropna()
+                ret = df.loc[:, [to, datacol]].dropna()
+                if to in ["ISO2", "ISO3"]:
+                    ret.loc[:, to] = (
+                        ret.loc[:, to]
+                        .str.split("|")
+                        .apply(lambda x: "".join((c for c in x[0] if c.isalnum())))
+                    )
+                return ret
 
             return fun_provided
 
@@ -606,6 +613,12 @@ def convert(
             else:
                 outlist[ind_names] = []
                 for etr in result_list:
+                    if to[0].lower() in ["iso2", "iso3"]:
+                        # remove regex characters from output
+                        etr = "".join(
+                            c for c in etr.split("|")[0] if c.isalnum()
+                        ).upper()
+
                     try:
                         conv_etr = int(etr)
                     except ValueError:

diff --git a/country_converter/country_data.tsv b/country_converter/country_data.tsv
@@ -239,7 +239,7 @@ Tuvalu	Tuvalu	tuvalu	TV	TUV	798	798	227	416	Oceania	Polynesia	WW	WA	WA	WWW	WWA	W
 Uganda	Republic of Uganda	uganda	UG	UGA	800	800	226	190	Africa	Eastern Africa	WW	WF	WF	WWW	WWF	WWF	RoW	UGA	AFR	Eastern Africa	SSA												1962	1962		RoW							Other Africa	285
 Ukraine	Ukraine	ukrain	UA	UKR	804	804	230	63	Europe	Eastern Europe	WW	WE	WE	WWW	WWE	WWE	RoW	UKR	FSU	Ukraine region	REF												1945	1945		RoW							Ukraine	85
 United Arab Emirates	United Arab Emirates	emirates|^u\.?a\.?e\.?$|united.?arab.?em	AE	ARE	784	784	225	156	Asia	Western Asia	WW	WM	WM	WWW	WWM	WWM	RoW	ARE	MEA	Middle East	MEA												1971	1971		RoW							United Arab Emirates	576
-United Kingdom	United Kingdom of Great Britain and Northern Ireland	.*(united.?kingdom|britain|^u\.?k\.?$|gb)	^uk$|^gb$	GBR	826	826	229	95	Europe	Northern Europe	GB	GB	GB	GBR	GBR	GBR	GBR	GBR	WEU	Western Europe	EUR	1961		EU28		EU27_2007	EU25	EU15	EU12	EEA			1945	1945		EU					G7	G20	United Kingdom	12
+United Kingdom	United Kingdom of Great Britain and Northern Ireland	.*(united.?kingdom|britain|^u\.?k\.?$|gb)	^GB$|^UK$	GBR	826	826	229	95	Europe	Northern Europe	GB	GB	GB	GBR	GBR	GBR	GBR	GBR	WEU	Western Europe	EUR	1961		EU28		EU27_2007	EU25	EU15	EU12	EEA			1945	1945		EU					G7	G20	United Kingdom	12
 United States	United States of America	^(?!.*islands).*united.?states|^u\.?s\.?a\.?$|^u\.?s\.?$	US	USA	840	840	231	102	America	Northern America	US	US	US	USA	USA	USA	USA	USA	NAM	USA	USA	1961											1945	1945		HI		APEC			G7	G20	United States	302
 United States Minor Outlying Islands	United States Minor Outlying Islands	minor.?outlying.?is	UM	UMI	581	581	232		Oceania	Micronesia	WW	WA	WA	WWW	WWA	WWA	RoW				OAS															RoW							Other non-OECD Asia	
 United States Virgin Islands	Virgin Islands of the United States	^(?=.*\bu\.?\s?s).*virgin|^(?=.*states).*virgin	VI	VIR	850	850	240	422	America	Caribbean	WW	WL	WL	WWW	WWL	WWL	RoW	VIR	NAM	Central America	LAM															RoW							Other non-OECD Americas	

diff --git a/country_converter/version.py b/country_converter/version.py
@@ -1 +1 @@
-__version__ = "0.7.6"
+__version__ = "0.7.7"
diff --git a/tests/test_functionality.py b/tests/test_functionality.py
@@ -128,6 +128,18 @@ def test_alternative_names(get_regex_test_data):
         )
 
 
+def test_toISO2_conversion():
+    converter = coco.CountryConverter()
+    assert "DE" == converter.convert("DEU", src="ISO3", to="ISO2")
+    assert "GB" == converter.convert("GBR", src="ISO3", to="ISO2")
+    assert "GB" == converter.convert("UK", src="ISO2", to="ISO2")
+    assert "GB" == converter.convert("UK", to="ISO2")
+    assert "GB" == converter.convert("GB", to="ISO2")
+    assert "GB" == converter.convert("GBR", to="ISO2")
+    assert "TR" == converter.convert("TR", src="ISO2", to="ISO2")
+    assert "TR" == converter.convert("TUR", src="ISO3", to="ISO2")
+
+
 def test_additional_country_file():
     converter_basic = coco.CountryConverter()
     converter_extended = coco.CountryConverter(additional_data=custom_data)
@@ -496,7 +508,7 @@ def test_EU_output():
     EU27_2007 = cc.EU27_2007as("ISO2")
     assert len(EU27_2007 == 27)
     assert cc.convert("Croatia", to="ISO2") not in EU27_2007.ISO2.tolist()
-    assert cc.convert("UK", src="regex", to="ISO2") in EU27_2007.ISO2.tolist()
+    assert cc.convert("GB", src="regex", to="ISO2") in EU27_2007.ISO2.tolist()
 
 
 def test_EXIO_output():