csv-detective 0.10.2.dev4__py3-none-any.whl → 0.10.3.dev1__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (7) hide show
  1. csv_detective/formats/date.py +10 -1
  2. {csv_detective-0.10.2.dev4.dist-info → csv_detective-0.10.3.dev1.dist-info}/METADATA +1 -1
  3. {csv_detective-0.10.2.dev4.dist-info → csv_detective-0.10.3.dev1.dist-info}/RECORD +7 -7
  4. {csv_detective-0.10.2.dev4.dist-info → csv_detective-0.10.3.dev1.dist-info}/WHEEL +0 -0
  5. {csv_detective-0.10.2.dev4.dist-info → csv_detective-0.10.3.dev1.dist-info}/entry_points.txt +0 -0
  6. {csv_detective-0.10.2.dev4.dist-info → csv_detective-0.10.3.dev1.dist-info}/licenses/LICENSE +0 -0
  7. {csv_detective-0.10.2.dev4.dist-info → csv_detective-0.10.3.dev1.dist-info}/top_level.txt +0 -0
csv_detective/formats/date.py CHANGED
@@ -57,7 +57,9 @@ string_month_pattern = (
57
57
 
58
58
 
59
59
  def _is(val):
60
- # early stops, to cut processing time
60
+ # many early stops, to cut processing time
61
+ # and avoid the costly use of date_casting as much as possible
62
+ # /!\ timestamps are considered ints, not dates
61
63
  if not isinstance(val, str) or len(val) > 20 or len(val) < 8:
62
64
  return False
63
65
  # if it's a usual date pattern
@@ -70,8 +72,13 @@ def _is(val):
70
72
  ]
71
73
  ):
72
74
  return True
75
+ if re.match(r"^-?\d+[\.|,]\d+$", val):
76
+ # regular floats are excluded
77
+ return False
78
+ # not enough digits => not a date (slightly arbitrary)
73
79
  if sum([char.isdigit() for char in val]) / len(val) < threshold:
74
80
  return False
81
+ # last resort
75
82
  res = date_casting(val)
76
83
  if not res or res.hour or res.minute or res.second:
77
84
  return False
@@ -86,6 +93,7 @@ _test_values = {
86
93
  "15 décembre 1985",
87
94
  "02 05 2003",
88
95
  "20030502",
96
+ "2003.05.02",
89
97
  "1993-12/02",
90
98
  ],
91
99
  False: [
@@ -96,5 +104,6 @@ _test_values = {
96
104
  "12152003",
97
105
  "20031512",
98
106
  "02052003",
107
+ "6.27367393749392839",
99
108
  ],
100
109
  }
{csv_detective-0.10.2.dev4.dist-info → csv_detective-0.10.3.dev1.dist-info}/METADATA RENAMED
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: csv-detective
3
- Version: 0.10.2.dev4
3
+ Version: 0.10.3.dev1
4
4
  Summary: Detect tabular files column content
5
5
  Author-email: "data.gouv.fr" <opendatateam@data.gouv.fr>
6
6
  License: MIT
{csv_detective-0.10.2.dev4.dist-info → csv_detective-0.10.3.dev1.dist-info}/RECORD RENAMED
@@ -28,7 +28,7 @@ csv_detective/formats/code_rna.py,sha256=o6Kptrux6T2bSnWHi7MBCqIfVKbMMeN4dHlxxzk
28
28
  csv_detective/formats/code_waldec.py,sha256=j4-xpj_73c7IdgLoZJY_kRVj3HkpB7RFfGPN4NwPmVo,303
29
29
  csv_detective/formats/commune.py,sha256=QVscVy5Ij9kdzKJgIG2aFC_v1IRsov5M9Zkj_SHDWgs,541
30
30
  csv_detective/formats/csp_insee.py,sha256=y1w9zPQvijQi5v1Cuye0aX87ZVDC4FeFx1YC0dLqqp8,688
31
- csv_detective/formats/date.py,sha256=Q6w1azLKNshJJVLOPBHj-77ZinXYMW_EKp_BGDshLLE,2802
31
+ csv_detective/formats/date.py,sha256=caMMvcqkbON8Cxp9oDYZsfmkSXuu-PiiJi8YUbypBso,3167
32
32
  csv_detective/formats/date_fr.py,sha256=YnNXSgT6QekfTUJoS5yuRX8LeK-fmVDgLgVP9cP0e4M,505
33
33
  csv_detective/formats/datetime_aware.py,sha256=izKo6CA-MNIzmmM3Br4-FOESyqCS_YYK8N4V9D6CVEI,1909
34
34
  csv_detective/formats/datetime_naive.py,sha256=DZ0apAm3vIy4cdm5DynAeRueI_8rhuHYQtAOZ5yyZ5k,1681
@@ -86,7 +86,7 @@ csv_detective/parsing/csv.py,sha256=5rw6gXZFQC1T4NT9CnW0AumidrYOkF8kjrfWGmk949I,
86
86
  csv_detective/parsing/excel.py,sha256=tb65I78tdYlZci_tzvvQt8U6bZSYKjeVdn2CEvsET1o,6972
87
87
  csv_detective/parsing/load.py,sha256=f-8aKiNpy_47qg4Lq-UZUR4NNrbJ_-KEGvcUQZ8cmb0,4317
88
88
  csv_detective/parsing/text.py,sha256=yDAcop5xJQc25UtbZcV0guHXAZQfm-H8WuJORTy8Rr8,1734
89
- csv_detective-0.10.2.dev4.dist-info/licenses/LICENSE,sha256=A1dQrzxyxRHRih02KwibWj1khQyF7GeA6SqdOU87Gk4,1088
89
+ csv_detective-0.10.3.dev1.dist-info/licenses/LICENSE,sha256=A1dQrzxyxRHRih02KwibWj1khQyF7GeA6SqdOU87Gk4,1088
90
90
  tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
91
91
  tests/test_example.py,sha256=uTWswvUzBWEADGXZmMAdZvKhKvIjvT5zWOVVABgCDN4,1987
92
92
  tests/test_fields.py,sha256=EuD2F1JUR8y88Hm-AYuJ5X7AKkGSyLIQfsGdxYgIWng,5680
@@ -104,8 +104,8 @@ tests/data/file.ods,sha256=4dR7zWptz5djALIBVeWHQ20GaZNfA63fevIJGFIk1_U,11832
104
104
  tests/data/file.xls,sha256=QYmNX3FF0QfcQSzYQMtaMJaepJf5EZpDa1miKc4wMdQ,21495
105
105
  tests/data/file.xlsx,sha256=naWzL02PK4pdIjMzfEyfSW9GQhkYYd_e7bpJvB8Pb2w,8314
106
106
  tests/data/xlsx_file,sha256=NyOyN_rIe7ryJuHQLqjxVdKCc8V4s5pxyHl6wWFykCM,8305
107
- csv_detective-0.10.2.dev4.dist-info/METADATA,sha256=JkKreIYD_a8Qgxt6LgKR3kxYaiDHVhSjiBm5WpP1_Zk,11082
108
- csv_detective-0.10.2.dev4.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
109
- csv_detective-0.10.2.dev4.dist-info/entry_points.txt,sha256=JjweTReFqKJmuvkegzlew2j3D5pZzfxvbEGOtGVGmaY,56
110
- csv_detective-0.10.2.dev4.dist-info/top_level.txt,sha256=KDI4gyOpkmormGgUvSWrE3jen2e0unIsxR2b96DRvcw,25
111
- csv_detective-0.10.2.dev4.dist-info/RECORD,,
107
+ csv_detective-0.10.3.dev1.dist-info/METADATA,sha256=gJcwR3wni376q4qqk8xOG6uy9W4fwA5enfmfT066990,11082
108
+ csv_detective-0.10.3.dev1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
109
+ csv_detective-0.10.3.dev1.dist-info/entry_points.txt,sha256=JjweTReFqKJmuvkegzlew2j3D5pZzfxvbEGOtGVGmaY,56
110
+ csv_detective-0.10.3.dev1.dist-info/top_level.txt,sha256=KDI4gyOpkmormGgUvSWrE3jen2e0unIsxR2b96DRvcw,25
111
+ csv_detective-0.10.3.dev1.dist-info/RECORD,,