hohn/sarif-cli
1
0
Fork 0
mirror of https://github.com/hohn/sarif-cli.git synced 2025-12-16 17:23:03 +01:00
Code Issues Packages Projects Releases Wiki Activity

sarif-extract-tables: include relatedLocations from both sources

Browse Source 
With the addition of the path-problem output, include both as sources (left joins)
for relatedLocations:

    pd.concat([sf(4055)[['relatedLocations', 'struct_id']],
              sf(9699)[['relatedLocations', 'struct_id']]])
This commit is contained in:
Michael Hohn
2022-02-22 17:35:39 -08:00
committed by =Michael Hohn
parent 1dbd240b5b
commit 9c151e295b
1 changed files with 36 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

66
bin/sarif-extract-tables
View File

@@ -140,36 +140,6 @@ d1 = (
d2 = (d1[['struct_id_4055', 'uri', 'startLine', 'startColumn', 'endLine', 'endColumn', 'text_4055']]
.rename({'text_4055': 'message'}, axis='columns'))
#
# Form the relatedLocation dataframe via joins. This is subtly different from d1:
# left_on=relatedLocations, and no left_on='message_4055'
dr1 = (
sf(4055)
.merge(af('0350'), how="left", left_on='relatedLocations', right_on='array_id', validate="1:m")
.drop(columns=['relatedLocations', 'array_id', 'value_index', 'type_at_index'])
#
.merge(sf(2683), how="left", left_on='id_or_value_at_index', right_on='struct_id',
suffixes=("_4055", "_2683"), validate="1:m")
.drop(columns=['struct_id_2683', 'id_or_value_at_index'])
.merge(sf(4963), how="left", left_on='physicalLocation', right_on='struct_id', validate="1:m")
#
.drop(columns=['struct_id', 'physicalLocation'])
.merge(sf(6299), how="left", left_on='region', right_on='struct_id', validate="1:m")
.drop(columns=['struct_id', 'region'])
.merge(sf(2685), how="left", left_on='artifactLocation', right_on='struct_id', validate="1:m")
#
.drop(columns=['struct_id', 'artifactLocation'])
.merge(sf(2774), how="left", left_on='message_2683', right_on='struct_id', validate="1:m")
.drop(columns=['struct_id', 'message_2683'])
)
# Keep columns of interest
dr2 = (dr1[['struct_id_4055', 'uri', 'startLine', 'startColumn', 'endLine', 'endColumn', 'text']]
.rename({'text': 'message'}, axis='columns'))
# Remove dummy locations previously injected by signature.fillsig
dr3 = dr2[dr2.uri != 'scli-dyys dummy value']
#
# Form the codeFlows dataframe
#
@@ -230,6 +200,42 @@ dco2 = (dco1[['uri',
# Remove dummy locations previously injected by signature.fillsig
dco3 = dco2[dco2.uri != 'scli-dyys dummy value']
#
# Form the relatedLocation dataframe via joins, starting from the union of
# relatedLocations from `kind problem` (sf(4055)) and `kind path-problem`
# (sf(9699)). This is only sligthly different from d1: left_on=relatedLocations,
# and no left_on='message_4055'
#
dr1 = (
pd.concat([sf(4055)[['relatedLocations', 'struct_id']], sf(9699)[['relatedLocations', 'struct_id']]])
.merge(af('0350'), how="left", left_on='relatedLocations', right_on='array_id', validate="1:m")
.drop(columns=['relatedLocations', 'array_id', 'value_index', 'type_at_index'])
#
.merge(sf(2683), how="left", left_on='id_or_value_at_index', right_on='struct_id',
suffixes=("_4055_9699", "_2683"), validate="1:m")
.drop(columns=['struct_id_2683', 'id_or_value_at_index'])
#
.merge(sf(4963), how="left", left_on='physicalLocation', right_on='struct_id', validate="1:m")
.drop(columns=['struct_id', 'physicalLocation'])
#
.merge(sf(6299), how="left", left_on='region', right_on='struct_id', validate="1:m")
.drop(columns=['struct_id', 'region'])
#
.merge(sf(2685), how="left", left_on='artifactLocation', right_on='struct_id', validate="1:m")
.drop(columns=['struct_id', 'artifactLocation'])
#
.merge(sf(2774), how="left", left_on='message', right_on='struct_id', validate="1:m")
.drop(columns=['struct_id', 'message'])
)
# Keep columns of interest
dr2 = (dr1[['struct_id_4055_9699', 'uri', 'startLine', 'startColumn', 'endLine', 'endColumn', 'text']]
.rename({'text': 'message', 'struct_id_4055_9699': 'struct_id'}, axis='columns'))
# Remove dummy locations previously injected by signature.fillsig
dr3 = dr2[dr2.uri != 'scli-dyys dummy value']
#
# Write output
#