Skip to content

Commit 66fcaa7

Browse files
committed
add alt mapping to microsoft_sentinel parser
1 parent e4602e0 commit 66fcaa7

File tree

4 files changed

+66
-9
lines changed

4 files changed

+66
-9
lines changed

uncoder-core/app/translator/core/mapping.py

Lines changed: 22 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -186,6 +186,28 @@ def get_source_mappings_by_fields_and_log_sources(
186186

187187
return by_log_sources_and_fields or by_fields or [self._source_mappings[DEFAULT_MAPPING_NAME]]
188188

189+
def get_alt_source_mappings_by_fields_and_log_sources(
190+
self, field_names: list[str], log_sources: dict[str, list[Union[int, str]]], alt_mapping: str
191+
) -> list[SourceMapping]:
192+
by_log_sources_and_fields = []
193+
by_fields = []
194+
for source_mapping in self._alternative_mappings.get(alt_mapping).values():
195+
if source_mapping.source_id == DEFAULT_MAPPING_NAME:
196+
continue
197+
198+
if source_mapping.fields_mapping.is_suitable(field_names):
199+
by_fields.append(source_mapping)
200+
201+
log_source_signature: LogSourceSignature = source_mapping.log_source_signature
202+
if log_source_signature and log_source_signature.is_suitable(**log_sources):
203+
by_log_sources_and_fields.append(source_mapping)
204+
205+
return (
206+
by_log_sources_and_fields
207+
or by_fields
208+
or [self._alternative_mappings.get(alt_mapping)[DEFAULT_MAPPING_NAME]]
209+
)
210+
189211
def get_source_mapping(self, source_id: str) -> Optional[SourceMapping]:
190212
return self._source_mappings.get(source_id)
191213

uncoder-core/app/translator/core/parser.py

Lines changed: 21 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -24,7 +24,7 @@
2424
from app.translator.core.exceptions.parser import TokenizerGeneralException
2525
from app.translator.core.functions import PlatformFunctions
2626
from app.translator.core.mapping import BasePlatformMappings, SourceMapping
27-
from app.translator.core.models.functions.base import Function
27+
from app.translator.core.models.functions.base import Function, ParsedFunctions
2828
from app.translator.core.models.platform_details import PlatformDetails
2929
from app.translator.core.models.query_container import RawQueryContainer, TokenizedQueryContainer
3030
from app.translator.core.models.query_tokens.field import Field
@@ -51,6 +51,9 @@ def parse_raw_query(self, text: str, language: str) -> RawQueryContainer:
5151
def parse(self, raw_query_container: RawQueryContainer) -> TokenizedQueryContainer:
5252
raise NotImplementedError("Abstract method")
5353

54+
def _parse_query(self, query: str) -> tuple[str, dict[str, Union[list[str], list[int]]], Optional[ParsedFunctions]]:
55+
raise NotImplementedError("Abstract method")
56+
5457

5558
class PlatformQueryParser(QueryParser, ABC):
5659
mappings: BasePlatformMappings = None
@@ -80,11 +83,24 @@ def get_field_tokens(
8083
return query_field_tokens, function_field_tokens, function_field_tokens_map
8184

8285
def get_source_mappings(
83-
self, field_tokens: list[Field], log_sources: dict[str, list[Union[int, str]]]
86+
self,
87+
field_tokens: list[Field],
88+
log_sources: dict[str, list[Union[int, str]]],
89+
alt_mapping: Optional[str] = None,
8490
) -> list[SourceMapping]:
8591
field_names = [field.source_name for field in field_tokens]
86-
source_mappings = self.mappings.get_source_mappings_by_fields_and_log_sources(
87-
field_names=field_names, log_sources=log_sources
88-
)
92+
if alt_mapping:
93+
source_mappings = self.mappings.get_alt_source_mappings_by_fields_and_log_sources(
94+
field_names=field_names, log_sources=log_sources, alt_mapping=alt_mapping
95+
)
96+
else:
97+
source_mappings = self.mappings.get_source_mappings_by_fields_and_log_sources(
98+
field_names=field_names, log_sources=log_sources
99+
)
89100
self.tokenizer.set_field_tokens_generic_names_map(field_tokens, source_mappings, self.mappings.default_mapping)
90101
return source_mappings
102+
103+
def get_source_mapping_ids_by_logsources(self, query: str) -> Optional[list[str]]:
104+
_, parsed_logsources, _ = self._parse_query(query=query)
105+
if parsed_logsources:
106+
return self.mappings.get_source_mappings_by_log_sources(parsed_logsources)

uncoder-core/app/translator/platforms/microsoft/parsers/microsoft_sentinel.py

Lines changed: 7 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -16,6 +16,7 @@
1616
-----------------------------------------------------------------
1717
"""
1818

19+
from typing import Optional, Union
1920

2021
from app.translator.core.models.functions.base import ParsedFunctions
2122
from app.translator.core.models.platform_details import PlatformDetails
@@ -37,7 +38,7 @@ class MicrosoftSentinelQueryParser(PlatformQueryParser):
3738

3839
wrapped_with_comment_pattern = r"^\s*//.*(?:\n|$)"
3940

40-
def _parse_query(self, query: str) -> tuple[str, dict[str, list[str]], ParsedFunctions]:
41+
def _parse_query(self, query: str) -> tuple[str, dict[str, Union[list[str], list[int]]], Optional[ParsedFunctions]]:
4142
table, query, functions = self.platform_functions.parse(query)
4243
log_sources = {"table": [table]}
4344
return query, log_sources, functions
@@ -48,7 +49,11 @@ def parse(self, raw_query_container: RawQueryContainer) -> TokenizedQueryContain
4849
query_field_tokens, function_field_tokens, function_field_tokens_map = self.get_field_tokens(
4950
query_tokens, functions.functions
5051
)
51-
source_mappings = self.get_source_mappings(query_field_tokens + function_field_tokens, log_sources)
52+
source_mappings = self.get_source_mappings(
53+
field_tokens = query_field_tokens + function_field_tokens,
54+
log_sources=log_sources,
55+
alt_mapping=raw_query_container.meta_info.source_alt_mapping
56+
)
5257
meta_info = raw_query_container.meta_info
5358
meta_info.query_fields = query_field_tokens
5459
meta_info.function_fields = function_field_tokens

uncoder-core/app/translator/translator.py

Lines changed: 16 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -105,7 +105,11 @@ def __translate_one(
105105
target_alt_mapping: Optional[str] = None,
106106
) -> (bool, str):
107107
status, parsed_data = self.__parse_incoming_data(
108-
text=text, source=source, target=target, source_alt_mapping=source_alt_mapping, target_alt_mapping=target_alt_mapping
108+
text=text,
109+
source=source,
110+
target=target,
111+
source_alt_mapping=source_alt_mapping,
112+
target_alt_mapping=target_alt_mapping,
109113
)
110114
if not status:
111115
return status, parsed_data
@@ -149,9 +153,19 @@ def translate_one(
149153
target_alt_mapping: Optional[str] = None,
150154
) -> (bool, str):
151155
if source == target:
156+
if target_alt_mapping or source_alt_mapping:
157+
message = (
158+
"Currently, Uncoder doesn't support translation between "
159+
"non-default data schemas of the same platform."
160+
)
161+
return False, message
152162
return True, text
153163
return self.__translate_one(
154-
text=text, source=source, target=target, source_alt_mapping=source_alt_mapping, target_alt_mapping=target_alt_mapping
164+
text=text,
165+
source=source,
166+
target=target,
167+
source_alt_mapping=source_alt_mapping,
168+
target_alt_mapping=target_alt_mapping,
155169
)
156170

157171
def translate_all(self, text: str, source: str) -> list[dict]:

0 commit comments

Comments
 (0)
pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy