csv-detective 0.9.3.dev2052__py3-none-any.whl → 0.9.3.dev2090__py3-none-any.whl

csv_detective/detect_fields/temp/datetime_aware/__init__.py

@@ -6,19 +6,20 @@ from csv_detective.detect_fields.temp.date import aaaammjj_pattern, date_casting
 PROPORTION = 1
 threshold = 0.7
 
-# matches AAAA-MM-JJTHH:MM:SS(.dddddd)±HH:MM with any of the listed separators for the date OR NO SEPARATOR
+# matches AAAA-MM-JJTHH:MM:SS(.dddddd)(±HH:MM|Z) with any of the listed separators for the date OR NO SEPARATOR
 pat = (
     aaaammjj_pattern.replace("$", "")
-    + r"(T|\s)(0\d|1[0-9]|2[0-3]):([0-5][0-9]):([0-5][0-9])(.\d{1,6})?[+-](0\d|1[0-9]|2[0-3]):([0-5][0-9])$"
+    + r"(T|\s)(0\d|1[0-9]|2[0-3]):([0-5][0-9]):([0-5][0-9])(.\d{1,6})"
+    + r"?(([+-](0\d|1[0-9]|2[0-3]):([0-5][0-9]))|Z)$"
 )
 
 
 def _is(val: Optional[Any]) -> bool:
     """Detects timezone-aware datetimes only"""
     # early stops, to cut processing time
-    # 21 is the minimal length of a datetime format YYMMDDTHH:MM:SS+HH:MM
+    # 16 is the minimal length of a datetime format YYMMDDTHH:MM:SSZ
     # 32 is the maximal length of an ISO datetime format YYYY-MM-DDTHH:MM:SS.dddddd+HH:MM, keeping some slack
-    if not isinstance(val, str) or len(val) > 35 or len(val) < 21:
+    if not isinstance(val, str) or len(val) > 35 or len(val) < 16:
         return False
     # if usual format, no need to parse
     if bool(re.match(pat, val)):

csv_detective/detect_fields/temp/datetime_naive/__init__.py

@@ -9,7 +9,7 @@ threshold = 0.7
 # matches AAAA-MM-JJTHH:MM:SS(.dddddd)Z with any of the listed separators for the date OR NO SEPARATOR
 pat = (
     aaaammjj_pattern.replace("$", "")
-    + r"(T|\s)(0\d|1[0-9]|2[0-3]):([0-5][0-9]):([0-5][0-9])(.\d{1,6})?Z?$"
+    + r"(T|\s)(0\d|1[0-9]|2[0-3]):([0-5][0-9]):([0-5][0-9])(.\d{1,6})?$"
 )
 
 

{csv_detective-0.9.3.dev2052.dist-info → csv_detective-0.9.3.dev2090.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: csv-detective
-Version: 0.9.3.dev2052
+Version: 0.9.3.dev2090
 Summary: Detect tabular files column content
 Author-email: Etalab <opendatateam@data.gouv.fr>
 License: MIT

{csv_detective-0.9.3.dev2052.dist-info → csv_detective-0.9.3.dev2090.dist-info}/RECORD

@@ -67,8 +67,8 @@ csv_detective/detect_fields/other/url/__init__.py,sha256=L7h9fZldh1w86XwCx0x3Q1T
 csv_detective/detect_fields/other/uuid/__init__.py,sha256=XFxbIsdIhRw0dtFxBXQBhicE4yy7P4jmwYXeJhq6FVY,215
 csv_detective/detect_fields/temp/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 csv_detective/detect_fields/temp/date/__init__.py,sha256=JtWaK8hkzBaIUc-fu0G7lIFpWqCfraRh6l0Mo65U3b0,2155
-csv_detective/detect_fields/temp/datetime_aware/__init__.py,sha256=ZDNUcbU0ZJzaxUt0Utc1Y9dRrq4HHW9uCbcnOuz5Sfk,1247
-csv_detective/detect_fields/temp/datetime_naive/__init__.py,sha256=U1mlQdbvEeJNZgMiYY7xv-_EIBRqzxHsKUgU3ZoF0FM,1088
+csv_detective/detect_fields/temp/datetime_aware/__init__.py,sha256=TV-Bden-3SFm9RPBtem1biqoxyEjsshZm9fFbnNXAIY,1263
+csv_detective/detect_fields/temp/datetime_naive/__init__.py,sha256=kodVtLa9ZBy7eJDNo2PQOdgoTXlXkVTUN8Vk03lG1n0,1086
 csv_detective/detect_fields/temp/datetime_rfc822/__init__.py,sha256=-pFdIIPgaLq2_QbFJ9zwy4YIwZuC73F0A_cNDntTuvQ,512
 csv_detective/detect_fields/temp/year/__init__.py,sha256=gHchVciZExbGZLMBcbBaDXB0IgGptkQc4RhfSOMY0Ww,194
 csv_detective/detect_labels/__init__.py,sha256=93s93DRNeFw9fJiGp0rW3iRWZX3WOeVau2PAaF4QlPE,1777
@@ -149,18 +149,18 @@ csv_detective/parsing/csv.py,sha256=fJkjKvyk7InkNnYKtmivyi48mmcwvrha7gvZ5J4-86A,
 csv_detective/parsing/excel.py,sha256=sKD5PRN1TlzPPOKFnZ3VRb0r1yIjPLlpxVWmZQeLYFk,7027
 csv_detective/parsing/load.py,sha256=C3M8nvgWenOb8aDFi5dpDGCoAw9EBqr4EB63zbz2M14,3699
 csv_detective/parsing/text.py,sha256=uz8wfmNTQnOd_4fjrIZ_5rxmFmgrg343hJh2szB73Hc,1770
-csv_detective-0.9.3.dev2052.dist-info/licenses/LICENSE,sha256=A1dQrzxyxRHRih02KwibWj1khQyF7GeA6SqdOU87Gk4,1088
+csv_detective-0.9.3.dev2090.dist-info/licenses/LICENSE,sha256=A1dQrzxyxRHRih02KwibWj1khQyF7GeA6SqdOU87Gk4,1088
 tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 tests/test_example.py,sha256=uTWswvUzBWEADGXZmMAdZvKhKvIjvT5zWOVVABgCDN4,1987
-tests/test_fields.py,sha256=IjUvDz155fl0XRw2ENEy5j_auyB3LTidXFi1bEWOCHg,13725
+tests/test_fields.py,sha256=R6r6dcUwPx9XWIoc1xH4z0HlCnTj_bmxw91H5Gfqq5I,13762
 tests/test_file.py,sha256=QEBv69P0bAKWBzhQ3KKOR1Z1RQSf5CVEilqBojwP2Yc,10791
 tests/test_labels.py,sha256=Y0XlOpztCyV65pk7iAS_nMMfdysoBujlBmz10vHul9A,469
 tests/test_structure.py,sha256=GRDYKy0UcdqlN4qglzsRC0puFj5cb-SVvONjvcPvtAA,1400
 tests/test_validation.py,sha256=ie-Xf0vk6-M6GQq-x7kY5yse1EmXfxQkbaV7fR3fvYo,3308
 venv/bin/activate_this.py,sha256=NRy3waFmwW1pOaNUp33wNN0vD1Kzkd-zXX-Sgl4EiVI,1286
 venv/bin/runxlrd.py,sha256=YlZMuycM_V_hzNt2yt3FyXPuwouMCmMhvj1oZaBeeuw,16092
-csv_detective-0.9.3.dev2052.dist-info/METADATA,sha256=Kn9PGSyDQEZfC9cFHyAB_nyyUupDB45IGnz2Ec2RktE,9735
-csv_detective-0.9.3.dev2052.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-csv_detective-0.9.3.dev2052.dist-info/entry_points.txt,sha256=JjweTReFqKJmuvkegzlew2j3D5pZzfxvbEGOtGVGmaY,56
-csv_detective-0.9.3.dev2052.dist-info/top_level.txt,sha256=cYKb4Ok3XgYA7rMDOYtxysjSJp_iUA9lJjynhVzue8g,30
-csv_detective-0.9.3.dev2052.dist-info/RECORD,,
+csv_detective-0.9.3.dev2090.dist-info/METADATA,sha256=nxz1XUFLoP5x9PTp8b0QHFOaH0Rz1mqITXd_zi-SbpY,9735
+csv_detective-0.9.3.dev2090.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+csv_detective-0.9.3.dev2090.dist-info/entry_points.txt,sha256=JjweTReFqKJmuvkegzlew2j3D5pZzfxvbEGOtGVGmaY,56
+csv_detective-0.9.3.dev2090.dist-info/top_level.txt,sha256=cYKb4Ok3XgYA7rMDOYtxysjSJp_iUA9lJjynhVzue8g,30
+csv_detective-0.9.3.dev2090.dist-info/RECORD,,

tests/test_fields.py

@@ -357,6 +357,7 @@ fields = {
         True: [
             "2021-06-22 10:20:10-04:00",
             "2030-06-22 00:00:00.0028+02:00",
+            "2000-12-21 10:20:10.1Z",
             "2024-12-19T10:53:36.428000+00:00",
             "1996/06/22 10:20:10 GMT",
         ],
@@ -365,7 +366,6 @@ fields = {
     datetime_naive: {
         True: [
             "2021-06-22 10:20:10",
-            "1999-12-01T00:00:00Z",
             "2030/06-22   00:00:00",
             "2030/06/22 00:00:00.0028",
         ],
@@ -373,6 +373,7 @@ fields = {
             "2021-06-22T30:20:10",
             "Sun, 06 Nov 1994 08:49:37 GMT",
             "2021-06-44 10:20:10+02:00",
+            "1999-12-01T00:00:00Z",
             "2021-06-44",
             "15 décembre 1985",
         ],
@@ -459,8 +460,8 @@ def test_priority(args):
         ("28/01/2000", date),
         ("2025-08-20T14:30:00+02:00", datetime_aware),
         ("2025/08/20 14:30:00.2763-12:00", datetime_aware),
-        ("1925_12_20T14:30:00.2763Z", datetime_naive),
-        ("1925 12 20 14:30:00Z", datetime_naive),
+        ("1925_12_20T14:30:00.2763", datetime_naive),
+        ("1925 12 20 14:30:00Z", datetime_aware),
     ),
 )
 def test_early_detection(args):