datasourcelib 0.1.13__py3-none-any.whl → 0.1.14__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.
- datasourcelib/utils/aggregation.py +8 -1
- {datasourcelib-0.1.13.dist-info → datasourcelib-0.1.14.dist-info}/METADATA +1 -1
- {datasourcelib-0.1.13.dist-info → datasourcelib-0.1.14.dist-info}/RECORD +6 -6
- {datasourcelib-0.1.13.dist-info → datasourcelib-0.1.14.dist-info}/WHEEL +1 -1
- {datasourcelib-0.1.13.dist-info → datasourcelib-0.1.14.dist-info}/licenses/LICENSE +0 -0
- {datasourcelib-0.1.13.dist-info → datasourcelib-0.1.14.dist-info}/top_level.txt +0 -0
|
@@ -2,6 +2,7 @@
|
|
|
2
2
|
import pandas as pd
|
|
3
3
|
from string import Formatter
|
|
4
4
|
from typing import Iterable, Any, Dict, List, Optional, Union
|
|
5
|
+
import regex as re
|
|
5
6
|
|
|
6
7
|
def _placeholders(fmt: str) -> List[str]:
|
|
7
8
|
"""
|
|
@@ -12,6 +13,12 @@ def _placeholders(fmt: str) -> List[str]:
|
|
|
12
13
|
def _safe_str(x) -> str:
|
|
13
14
|
return "" if pd.isna(x) else str(x).strip()
|
|
14
15
|
|
|
16
|
+
@staticmethod
|
|
17
|
+
def sanitize(s: str) -> str:
|
|
18
|
+
"""Keep only A-Z a-z 0-9 underscore/dash/equals in a safe way."""
|
|
19
|
+
# using the `regex` import already present as `re`
|
|
20
|
+
return re.sub(r'[^A-Za-z0-9_\-=]', '', s)
|
|
21
|
+
|
|
15
22
|
def generate_grouped_summaries(
|
|
16
23
|
df: pd.DataFrame,
|
|
17
24
|
aggregation_field: str,
|
|
@@ -146,7 +153,7 @@ def generate_grouped_summaries(
|
|
|
146
153
|
|
|
147
154
|
content = header + " " + " ".join(lines)
|
|
148
155
|
summaries.append(
|
|
149
|
-
{"content" : content, "id": group_value}
|
|
156
|
+
{"content" : content, "id": sanitize(group_value.replace(" ", "_").strip())}
|
|
150
157
|
)
|
|
151
158
|
|
|
152
159
|
return summaries
|
|
@@ -24,14 +24,14 @@ datasourcelib/strategies/incremental_load.py,sha256=CY1tAyXwjZLoq5zMLwB5i5qmT_L8
|
|
|
24
24
|
datasourcelib/strategies/ondemand_load.py,sha256=MgenKJbJePLeErdEkXKsz1h7RuR8yT0RV_X523G7UUs,1304
|
|
25
25
|
datasourcelib/strategies/timerange_load.py,sha256=W_sSZg059Lw2o9tmdGKM9D5-z1pph7AN1ftalXhuyjo,1557
|
|
26
26
|
datasourcelib/utils/__init__.py,sha256=9pSIpaK-kdmNuDzwl0Z7QU-_lV3cZE-iwOEPh3RBBTs,298
|
|
27
|
-
datasourcelib/utils/aggregation.py,sha256=
|
|
27
|
+
datasourcelib/utils/aggregation.py,sha256=_XzTxdGIc-nc0w1FE1NfPA6J1PmAKiSpz0sYU7yEU6s,6337
|
|
28
28
|
datasourcelib/utils/byte_reader.py,sha256=GaoPXwJa2YTWG1Kim0K6JG20eVSaWkZJd1o9bswxHmc,9082
|
|
29
29
|
datasourcelib/utils/exceptions.py,sha256=mgcDaW1k3VndgpMOwSm7NqgyRTvvE2a5ehn3x4fYQww,369
|
|
30
30
|
datasourcelib/utils/file_reader.py,sha256=Zr0rwNTRWE6KeVJEXgTOPS1_JI74LiUSiX5-6qojmN0,7301
|
|
31
31
|
datasourcelib/utils/logger.py,sha256=Sl6lNlvubxtK9ztzyq7vjGVyA8_-pZ_ixpk5jfVsh6U,424
|
|
32
32
|
datasourcelib/utils/validators.py,sha256=fLgmRAb5OZSdMVlHu_n0RKJUDl-G8dI8JsRSfxIquh8,205
|
|
33
|
-
datasourcelib-0.1.
|
|
34
|
-
datasourcelib-0.1.
|
|
35
|
-
datasourcelib-0.1.
|
|
36
|
-
datasourcelib-0.1.
|
|
37
|
-
datasourcelib-0.1.
|
|
33
|
+
datasourcelib-0.1.14.dist-info/licenses/LICENSE,sha256=9S0AcKETmp9XOcC73jEjN7WSkuSWGFGreiBat6ONClo,1087
|
|
34
|
+
datasourcelib-0.1.14.dist-info/METADATA,sha256=EyIP4yk74vb4-yhOWh--E9aXo8E1gt8RDVkYVoexfNI,1200
|
|
35
|
+
datasourcelib-0.1.14.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
|
|
36
|
+
datasourcelib-0.1.14.dist-info/top_level.txt,sha256=wIwiwdIj8T9pAvE2TkGLUvT2oIi43C2vkkTKibUlv3U,14
|
|
37
|
+
datasourcelib-0.1.14.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|