From b3f144696cdd5f21e6f53aa2571a04a55029b614 Mon Sep 17 00:00:00 2001
From: Daniel <d.hornung@indiscale.com>
Date: Wed, 29 May 2024 12:02:00 +0200
Subject: [PATCH] FIX: Better ID pattern, refactored a bit, added tests.
---
src/caoscrawler/converters.py | 8 ++++----
unittests/test_issues.py | 33 ++++++++++++++++++++-------------
2 files changed, 24 insertions(+), 17 deletions(-)
diff --git a/src/caoscrawler/converters.py b/src/caoscrawler/converters.py
index cce5146e..1a7f4d56 100644
--- a/src/caoscrawler/converters.py
+++ b/src/caoscrawler/converters.py
@@ -53,16 +53,16 @@ from .utils import has_parent
# by the converters:
SPECIAL_PROPERTIES = ("description", "name", "id", "path",
"file", "checksum", "size")
-# This RE is used to check whether a value is just a single variable:
-SINGLE_VAR_RE = re.compile(r"^\$(\{)?(?P<varname>\w+)(\})?$")
+ID_PATTERN = r"\D[.\w]*"
+SINGLE_VAR_RE = re.compile(r"^\$(\{)?(?P<varname>" + ID_PATTERN + r")(\})?$")
logger = logging.getLogger(__name__)
class CrawlerTemplate(Template):
- # This adds a dot to the default pattern.
+ # This also adds a dot to the default pattern.
# See: https://docs.python.org/3/library/string.html#template-strings
# Default flags is re.IGNORECASE
- braceidpattern = r"(?a:[_a-z][_\.a-z0-9]*)"
+ braceidpattern = ID_PATTERN
def _only_max(children_with_keys):
diff --git a/unittests/test_issues.py b/unittests/test_issues.py
index efc170ed..e6fe06ef 100644
--- a/unittests/test_issues.py
+++ b/unittests/test_issues.py
@@ -116,12 +116,15 @@ def test_issue_93():
cfood.yaml does not allow umlaut in $expression"""
values = GeneralStore()
expressions = [
- "1",
+ "foo",
+ "foo.bär",
+ "_1",
"Ä",
"ųøîµ",
]
for exp in expressions:
values[exp] = f"This is {exp}"
+ # ## Test preliminary check
# With braces
for exp in expressions:
assert replace_variables(f"${{{exp}}}", values) == f"This is {exp}"
@@ -129,15 +132,19 @@ def test_issue_93():
for exp in expressions:
assert replace_variables(f"${exp}", values) == f"This is {exp}"
-
-def test_crawler_template():
- temp = CrawlerTemplate("$bla")
- assert temp.safe_substitute(**{"bla": "test"}) == "test"
-
- # Umlauts are not replaced, because they are not contained
- # in the CrawlerTemplate's braceidpattern:
- temp = CrawlerTemplate("$blä")
- assert temp.safe_substitute(**{"blä": "test"}) == "$blä"
-
- temp = CrawlerTemplate("${blä}")
- assert temp.safe_substitute(**{"blä": "test"}) == "${blä}"
+ # ## Test actual replacement
+ for exp in expressions:
+ # as-is
+ propvalue = f"${{{exp}}}"
+ propvalue_template = CrawlerTemplate(propvalue)
+ # from IPython import embed
+ # embed()
+
+ assert propvalue_template.safe_substitute(**values.get_storage()) == f"This is {exp}"
+
+ # String embedded into context
+ propvalue = f"some text before >> ${{{exp}}} << some text after"
+ print(propvalue)
+ propvalue_template = CrawlerTemplate(propvalue)
+ assert (propvalue_template.safe_substitute(**values.get_storage())
+ == f"some text before >> This is {exp} << some text after")
--
GitLab