digital-forensics-labs/mobile-pii-discovery-agent
1
0
Fork 0
mirror of https://github.com/frankwxu/mobile-pii-discovery-agent.git synced 2026-02-20 13:40:41 +00:00
Code Issues Packages Projects Releases Wiki Activity

add postal address to config

Browse Source
This commit is contained in:
Frank Xu
2026-01-26 22:29:05 -05:00
parent fe649183a0
commit 1b3d3e2e95
51 changed files with 115 additions and 28 deletions
Download Patch File Download Diff File Expand all files Collapse all files

75
my_run_config.py
View File

@@ -1,31 +1,31 @@
db_files = [
# "test2.db",
# "users.db",
# "A1_commerce.db",
# "A1_msgstore.db",
# "A1_wa.db",
# "A2_core.db",
# "A2_journal.db",
# "A2_main.db",
# "A3_account1cache4.db",
# "A3_account2cache4.db",
# "A3_account3cache4.db",
# "A4_gmm_myplaces.db",
# "A4_gmm_storage.db",
# "A4_peopleCache_sharononeil368@gmail.com_com.google_14.db",
# "A5_SBrowser.db",
# "A5_SBrowser2.db",
# "A5_searchengine.db",
# "I1_CallHistory.sqlite",
# "I1_ChatStorage.sqlite",
# "I1_ContactsV2.sqlite",
# "I2_AddressBook.sqlitedb",
# "I2_AddressBookImages.sqlitedb",
# "I3_sms.db",
# "I4_CloudTabs.db",
# "I4_History.db",
# "I5_Calendar.sqlitedb",
# "I5_Extras.db",
"A1_commerce.db",
"A1_msgstore.db",
"A1_wa.db",
"A2_core.db",
"A2_journal.db",
"A2_main.db",
"A3_account1cache4.db",
"A3_account2cache4.db",
"A3_account3cache4.db",
"A4_gmm_myplaces.db",
"A4_gmm_storage.db",
"A4_peopleCache_sharononeil368@gmail.com_com.google_14.db",
"A5_SBrowser.db",
"A5_SBrowser2.db",
"A5_searchengine.db",
"I1_CallHistory.sqlite",
"I1_ChatStorage.sqlite",
"I1_ContactsV2.sqlite",
"I2_AddressBook.sqlitedb",
"I2_AddressBookImages.sqlitedb",
"I3_sms.db",
"I4_CloudTabs.db",
"I4_History.db",
"I5_Calendar.sqlitedb",
"I5_Extras.db",
]
PII_CONFIG = {
@@ -42,12 +42,31 @@ PII_CONFIG = {
"USERNAME": {
"type":"username",
"regex": r"\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\b",
"desc": "a username (also called a login name, user ID, or account name) is a unique string of characters that identifies a user on a computer system, website, application, or online platform"
"desc": "a username (also called a login name, user ID, or account name) is a unique string of characters that identifies a user on a computer system, website, application, or online platform. It is often created by the user during the registration process and is used in combination with a password to authenticate the user's identity"
},
"PERSON_NAME": {
"type":"person's name",
"type":"person name",
"regex": r"[A-Za-z][A-Za-z\s\.\-]{1,50}",
"desc": "a loosely structured human name-like strings that typically consist of a first name, a first name and a last name, and may also include middle names, initials, prefixes (e.g., Mr., Dr.), and suffixes (e.g., Jr., Sr.)"
}
},
"POSTAL_ADDRESS": {
"type": "US postal address",
# MAX RECALL prefilter (street number optional).
# Matches either:
# (1) PO Box patterns, OR
# (2) optional street number + some tokens + a street suffix, OR
# (3) street suffix with nearby tokens (even without a number).
"regex": r"(?is)\b("
r"(?:P\.?\s*O\.?\s*BOX|POST\s+OFFICE\s+BOX)\s*\d{1,6}"
r"|"
r"(?:\d{1,7}\s*)?" # OPTIONAL street number
r"(?:[A-Z0-9][A-Z0-9'.,/#\-]*\s*){1,25}?" # optional-ish tokens before suffix
r"(?:AVE|AVENUE|ST|STREET|RD|ROAD|BLVD|BOULEVARD|DR|DRIVE|LN|LANE|CT|COURT|PL|PLACE|WAY|"
r"PKWY|PARKWAY|CIR|CIRCLE|TER|TERRACE|HWY|HIGHWAY|TRL|TRAIL|SQ|SQUARE|PIKE|LOOP|RUN|WALK|PATH|BYP|BYPASS)\b"
r"(?:\s*(?:,|\s)\s*(?:N|S|E|W|NE|NW|SE|SW))?" # optional directional
r"(?:.{0,60}?\b\d{5}(?:-\d{4})?\b)?" # optional ZIP nearby
r")\b",
"desc": "a US postal address is a street-level mailing location in the United States, commonly appearing as a street name and suffix (e.g., 'Market St') optionally with a street number (e.g., '1500 Market St'), unit, city/state, ZIP, or a PO Box (e.g., 'P.O. Box 123')"
}
}