kodexa 7.0.12396563598__tar.gz → 7.0.12397101638__tar.gz
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.
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/PKG-INFO +1 -1
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/model.py +10 -15
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/ast.py +24 -18
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/pyproject.toml +1 -1
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/LICENSE +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/README.md +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/assistant/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/assistant/assistant.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/connectors/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/connectors/connectors.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/dataclasses/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/dataclasses/templates/llm_data_class.j2 +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/base.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/entities/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/entities/check_response.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/entities/product.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/entities/product_subscription.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/objects.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/persistence.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/utils.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/pipeline/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/pipeline/pipeline.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/platform/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/platform/client.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/platform/interaction.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/platform/kodexa.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/core.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/lexrules.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/lextab.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/lextab.pyi +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/parserules.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/parserules.pyi +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/parsetab.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/selectors/parsetab.pyi +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/spatial/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/spatial/azure_models.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/spatial/bbox_common.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/spatial/table_form_common.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/steps/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/steps/common.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/testing/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/testing/test_components.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/testing/test_utils.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/training/__init__.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/training/train_utils.py +0 -0
- {kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/utils/__init__.py +0 -0
@@ -293,8 +293,6 @@ class ContentNode(object):
|
|
293
293
|
if content is not None and len(self.get_content_parts()) == 0:
|
294
294
|
self.set_content_parts([content])
|
295
295
|
|
296
|
-
self.cached_all_content = None
|
297
|
-
|
298
296
|
def get_content_parts(self):
|
299
297
|
return self.document.get_persistence().get_content_parts(self)
|
300
298
|
|
@@ -793,12 +791,13 @@ class ContentNode(object):
|
|
793
791
|
result = self.select(selector, variables)
|
794
792
|
return result[0] if len(result) > 0 else None
|
795
793
|
|
796
|
-
def select(self, selector, variables=None):
|
794
|
+
def select(self, selector, variables=None, first_only=False):
|
797
795
|
"""Select and return the child nodes of this node that match the selector value.
|
798
796
|
|
799
797
|
Args:
|
800
798
|
selector (str): The selector (ie. //*)
|
801
799
|
variables (dict, optional): A dictionary of variable name/value to use in substituion; defaults to None. Dictionary keys should match a variable specified in the selector.
|
800
|
+
first_only (bool, optional): If True, only the first matching node will be returned; defaults to False.
|
802
801
|
|
803
802
|
Returns:
|
804
803
|
list[ContentNode]: A list of the matching content nodes. If no matches are found, the list will be empty.
|
@@ -815,18 +814,17 @@ class ContentNode(object):
|
|
815
814
|
from kodexa.selectors import parse
|
816
815
|
from kodexa.selectors.ast import SelectorContext
|
817
816
|
|
818
|
-
context = SelectorContext(self.document)
|
817
|
+
context = SelectorContext(self.document, first_only=first_only)
|
819
818
|
parsed_selector = parse(selector)
|
820
819
|
self.document.get_persistence().flush_cache()
|
821
820
|
return parsed_selector.resolve(self, variables, context)
|
822
821
|
|
823
|
-
def get_all_content(self, separator=" ", strip=True
|
822
|
+
def get_all_content(self, separator=" ", strip=True):
|
824
823
|
"""Get this node's content, concatenated with all of its children's content.
|
825
824
|
|
826
825
|
Args:
|
827
826
|
separator(str, optional): The separator to use in joining content together; defaults to " ".
|
828
827
|
strip(boolean, optional): Strip the result
|
829
|
-
use_cache(boolean, optional): Use the cache
|
830
828
|
|
831
829
|
Returns:
|
832
830
|
str: The complete content for this node concatenated with the content of all child nodes.
|
@@ -836,9 +834,6 @@ class ContentNode(object):
|
|
836
834
|
"This string is made up of multiple nodes"
|
837
835
|
"""
|
838
836
|
s = ""
|
839
|
-
if self.cached_all_content is not None and use_cache:
|
840
|
-
return self.cached_all_content
|
841
|
-
|
842
837
|
children = self.get_content_parts()
|
843
838
|
for part in children:
|
844
839
|
if isinstance(part, str):
|
@@ -861,14 +856,13 @@ class ContentNode(object):
|
|
861
856
|
s += separator
|
862
857
|
s += child.get_all_content(separator, strip=strip)
|
863
858
|
|
864
|
-
|
865
|
-
return self.cached_all_content
|
859
|
+
return s.strip() if strip else s
|
866
860
|
|
867
861
|
def adopt_children(self, nodes_to_adopt, replace=False):
|
868
862
|
"""This will take a list of content nodes and adopt them under this node, ensuring they are re-parented.
|
869
863
|
|
870
864
|
Args:
|
871
|
-
|
865
|
+
nodes_to_adopt (List[ContentNode]): A list of ContentNodes that will be added to the end of this node's children collection
|
872
866
|
replace (bool): If True, will remove all current children and replace them with the new list; defaults to True
|
873
867
|
|
874
868
|
>>> # select all nodes of type 'line', then the root node 'adopts' them
|
@@ -3109,17 +3103,18 @@ class Document(object):
|
|
3109
3103
|
>>> document.get_root().select_first('//*[hasTag($tagName)]', {"tagName": "div"})
|
3110
3104
|
ContentNode
|
3111
3105
|
"""
|
3112
|
-
result = self.select(selector, variables)
|
3106
|
+
result = self.select(selector, variables, first_only=True)
|
3113
3107
|
return result[0] if len(result) > 0 else None
|
3114
3108
|
|
3115
3109
|
def select(
|
3116
|
-
self, selector: str, variables: Optional[dict] = None
|
3110
|
+
self, selector: str, variables: Optional[dict] = None, first_only=False
|
3117
3111
|
) -> List[ContentNode]:
|
3118
3112
|
"""Execute a selector on the root node and then return a list of the matching nodes.
|
3119
3113
|
|
3120
3114
|
Args:
|
3121
3115
|
selector (str): The selector (ie. //*)
|
3122
3116
|
variables (Optional[dict): A dictionary of variable name/value to use in substituion; defaults to an empty
|
3117
|
+
first_only (bool): If True, only the first matching node is returned; defaults to False.
|
3123
3118
|
dictionary. Dictionary keys should match a variable specified in the selector.
|
3124
3119
|
|
3125
3120
|
Returns:
|
@@ -3131,7 +3126,7 @@ class Document(object):
|
|
3131
3126
|
if variables is None:
|
3132
3127
|
variables = {}
|
3133
3128
|
if self.content_node:
|
3134
|
-
result = self.content_node.select(selector, variables)
|
3129
|
+
result = self.content_node.select(selector, variables, first_only)
|
3135
3130
|
if isinstance(result, list):
|
3136
3131
|
return result
|
3137
3132
|
|
@@ -36,11 +36,12 @@ __all__ = [
|
|
36
36
|
|
37
37
|
|
38
38
|
class SelectorContext:
|
39
|
-
def __init__(self, document: Document):
|
39
|
+
def __init__(self, document: Document, first_only=False):
|
40
40
|
self.pattern_cache = {}
|
41
41
|
self.last_op = None
|
42
42
|
self.document: Document = document
|
43
43
|
self.stream = 0
|
44
|
+
self.first_only = first_only
|
44
45
|
|
45
46
|
def cache_pattern(self, pattern):
|
46
47
|
if pattern not in self.pattern_cache:
|
@@ -53,20 +54,26 @@ class PipelineExpression(object):
|
|
53
54
|
|
54
55
|
def __init__(self, left, op, right):
|
55
56
|
self.left = left
|
56
|
-
"""the left side of the pipeline expression"""
|
57
57
|
self.op = op
|
58
|
-
"""the operator of the pipeline expression"""
|
59
58
|
self.right = right
|
60
|
-
"""the right side of the pipeline expression"""
|
61
59
|
|
62
60
|
def resolve(self, content_node: ContentNode, variables, context: SelectorContext):
|
63
61
|
left_nodes = self.left.resolve(content_node, variables, context)
|
64
62
|
result_nodes: List[ContentNode] = []
|
65
63
|
context.stream = context.stream + 1
|
66
|
-
|
67
|
-
|
64
|
+
|
65
|
+
# If first_only is True and we already have left nodes, only process the first one
|
66
|
+
nodes_to_process = left_nodes[:1] if context.first_only and left_nodes else left_nodes
|
67
|
+
|
68
|
+
for node in nodes_to_process:
|
69
|
+
right_results = self.right.resolve(node, variables, context)
|
70
|
+
result_nodes.extend(right_results)
|
71
|
+
# If first_only is True and we found a match, return immediately
|
72
|
+
if context.first_only and result_nodes:
|
73
|
+
break
|
74
|
+
|
68
75
|
context.stream = context.stream - 1
|
69
|
-
return result_nodes
|
76
|
+
return result_nodes[:1] if context.first_only else result_nodes
|
70
77
|
|
71
78
|
|
72
79
|
class UnaryExpression(object):
|
@@ -181,17 +188,12 @@ class AbsolutePath(object):
|
|
181
188
|
|
182
189
|
|
183
190
|
class Step(object):
|
184
|
-
"""
|
185
|
-
A single step in a relative path. a; @b; text(); parent::foo:bar[5].
|
186
|
-
"""
|
191
|
+
"""A single step in a relative path."""
|
187
192
|
|
188
193
|
def __init__(self, axis, node_test, predicates):
|
189
194
|
self.axis = axis
|
190
|
-
"""the step's axis, or @ or None if abbreviated or undefined"""
|
191
195
|
self.node_test = node_test
|
192
|
-
"""a NameTest or NodeType object describing the test represented"""
|
193
196
|
self.predicates = predicates
|
194
|
-
"""a list of predicates filtering the step"""
|
195
197
|
|
196
198
|
def resolve(self, obj, variables, context: SelectorContext):
|
197
199
|
match = True
|
@@ -217,8 +219,9 @@ class Step(object):
|
|
217
219
|
return []
|
218
220
|
|
219
221
|
nodes = self.node_test.test(axis_node, variables, context)
|
220
|
-
|
221
222
|
final_nodes = []
|
223
|
+
|
224
|
+
# If first_only is True, only process until we find the first match
|
222
225
|
for node in nodes:
|
223
226
|
match = True
|
224
227
|
for predicate in self.predicates:
|
@@ -230,6 +233,8 @@ class Step(object):
|
|
230
233
|
|
231
234
|
if match:
|
232
235
|
final_nodes.append(node)
|
236
|
+
if context.first_only:
|
237
|
+
break
|
233
238
|
|
234
239
|
return final_nodes
|
235
240
|
|
@@ -247,9 +252,7 @@ class NameTest(object):
|
|
247
252
|
|
248
253
|
def __init__(self, prefix, name):
|
249
254
|
self.prefix = prefix
|
250
|
-
"""the namespace prefix used for the test, or None if unset"""
|
251
255
|
self.name = name
|
252
|
-
"""the node name used for the test, or *"""
|
253
256
|
|
254
257
|
def test(self, obj, variables, context: SelectorContext):
|
255
258
|
if isinstance(obj, ContentNode):
|
@@ -257,12 +260,15 @@ class NameTest(object):
|
|
257
260
|
if self.name == "*" or self.name == obj.node_type:
|
258
261
|
return [obj]
|
259
262
|
else:
|
260
|
-
|
263
|
+
nodes = context.document.get_persistence().get_content_nodes(
|
261
264
|
self.name, obj, context.last_op != "/"
|
262
265
|
)
|
266
|
+
# If first_only is True, return only the first matching node
|
267
|
+
return nodes[:1] if context.first_only else nodes
|
268
|
+
|
263
269
|
if isinstance(obj, ContentFeature):
|
264
270
|
return self.name == "*" or (
|
265
|
-
|
271
|
+
obj.feature_type == self.prefix and obj.name == self.name
|
266
272
|
)
|
267
273
|
return False
|
268
274
|
|
@@ -1,6 +1,6 @@
|
|
1
1
|
[tool.poetry]
|
2
2
|
name = "kodexa"
|
3
|
-
version = "7.0.
|
3
|
+
version = "7.0.012397101638"
|
4
4
|
description = "Python SDK for the Kodexa Platform"
|
5
5
|
authors = ["Austin Redenbaugh <austin@kodexa.com>", "Philip Dodds <philip@kodexa.com>", "Romar Cablao <rcablao@kodexa.com>", "Amadea Paula Dodds <amadeapaula@kodexa.com>"]
|
6
6
|
readme = "README.md"
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
{kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/dataclasses/templates/llm_data_class.j2
RENAMED
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
{kodexa-7.0.12396563598 → kodexa-7.0.12397101638}/kodexa/model/entities/product_subscription.py
RENAMED
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|