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
Files
59432cff2b08db783d926db9f3ec7bff37fb0753
mobile-pii-discovery-agent/model_PII_results/gpt4o/PII_A1_wa_20260131T203943Z.jsonl

6 lines
26 KiB
JSON
Raw Blame History

This file contains invisible Unicode characters
This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
{"db_path": "selectedDBs\\A1_wa.db", "PII_type": "EMAIL", "PII": [], "Num_of_PII": 0, "source_columns": [], "Raw_rows_first_100": [["12014946184@s.whatsapp.net"], ["12025692832@s.whatsapp.net"], ["12028177932@s.whatsapp.net"], ["12028275725@s.whatsapp.net"], ["120363098389106519@g.us"], ["12037188989@s.whatsapp.net"], ["12065937224@s.whatsapp.net"], ["12088549831@s.whatsapp.net"], ["12089234440@s.whatsapp.net"], ["12092759604@s.whatsapp.net"]], "Total_raw_rows": 10, "Exploration_sql": "SELECT email FROM wa_biz_profiles WHERE email REGEXP '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}' \nUNION ALL \nSELECT jid FROM wa_contacts WHERE jid REGEXP '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}' \nUNION ALL \nSELECT account_id FROM wa_biz_profiles_linked_accounts_table WHERE account_id REGEXP '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}' \nUNION ALL \nSELECT websites FROM wa_biz_profiles_websites WHERE websites REGEXP '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}';", "Extraction_sql": null, "PII_Prompt": "a unique identifier for a destination to which electronic mail (email) can be sent and received over the internet; examples include jane.doe@example.com, john.smith@provider.net, dev-team@startup.io, and user.name+label@domain.org"}
{"db_path": "selectedDBs\\A1_wa.db", "PII_type": "PHONE", "PII": ["19199037779", "8085096467", "5713349815", "16263678865", "16106046786", "7034241981", "5715917168", "2065937224", "5713298742", "8624338328", "18056377243", "2028177932", "2025692832", "9735203731", "81367430271", "7423794330", "5713298742", "5713349815", "8085096467", "8624338328", "7034241981", "2028177932", "5715917168", "2025692832", "16106046786", "16263678865", "2065937224", "19199037779", "18056377243", "9735203731", "81367430271", "7423794330"], "Num_of_PII": 32, "source_columns": ["wa_address_book.number", "wa_contacts.number"], "Raw_rows_first_100": [["+19199037779"], ["8085096467"], ["5713349815"], ["+16263678865"], ["+16106046786"], ["7034241981"], ["5715917168"], ["+12065937224"], ["5713298742"], ["8624338328"], ["+18056377243"], ["2028177932"], ["2025692832"], ["+19735203731"], ["+81367430271"], ["+17423794330"], ["5713298742"], ["5713349815"], ["8085096467"], ["8624338328"], ["7034241981"], ["2028177932"], ["5715917168"], ["2025692832"], ["+16106046786"], ["+16263678865"], ["+12065937224"], ["+19199037779"], ["+18056377243"], ["+19735203731"], ["+81367430271"], ["+17423794330"]], "Total_raw_rows": 32, "Exploration_sql": "SELECT number FROM wa_address_book WHERE number REGEXP '\\+?[0-9]{1,4}[- .]?\\(?[0-9]{1,3}?\\)?[- .]?[0-9]{1,4}[- .]?[0-9]{1,4}[- .]?[0-9]{1,9}' \nUNION ALL \nSELECT number FROM wa_biz_profiles WHERE number REGEXP '\\+?[0-9]{1,4}[- .]?\\(?[0-9]{1,3}?\\)?[- .]?[0-9]{1,4}[- .]?[0-9]{1,4}[- .]?[0-9]{1,9}' \nUNION ALL \nSELECT number FROM wa_contacts WHERE number REGEXP '\\+?[0-9]{1,4}[- .]?\\(?[0-9]{1,3}?\\)?[- .]?[0-9]{1,4}[- .]?[0-9]{1,4}[- .]?[0-9]{1,9}';", "Extraction_sql": "SELECT number FROM wa_address_book WHERE number REGEXP '\\+?[0-9]{1,4}[- .]?\\(?[0-9]{1,3}?\\)?[- .]?[0-9]{1,4}[- .]?[0-9]{1,4}[- .]?[0-9]{1,9}'\nUNION ALL\nSELECT number FROM wa_biz_profiles WHERE number REGEXP '\\+?[0-9]{1,4}[- .]?\\(?[0-9]{1,3}?\\)?[- .]?[0-9]{1,4}[- .]?[0-9]{1,4}[- .]?[0-9]{1,9}'\nUNION ALL\nSELECT number FROM wa_contacts WHERE number REGEXP '\\+?[0-9]{1,4}[- .]?\\(?[0-9]{1,3}?\\)?[- .]?[0-9]{1,4}[- .]?[0-9]{1,4}[- .]?[0-9]{1,9}';", "PII_Prompt": "a US phone number is a 10-digit NANP number (area code + exchange + line) that may be written as 2023133725, 202-313-3725, (202) 313-3725, 202.313.3725, +1 202 313 3725, or 1-202-313-3725"}
{"db_path": "selectedDBs\\A1_wa.db", "PII_type": "USERNAME", "PII": ["100674318659756", "103023648964858", "103929853546678", "103981426733067", "104569854005296", "106940592033823", "107967072444508", "108272048668771", "109766730866918", "110638642757646", "112103243395097", "11356128460981", "115504622653502", "115728095129747", "116032903606475", "12014946184", "12025692832", "12028177932", "12028275725", "120363098389106519", "120363144038483540", "120363159552753674", "120363160757210405", "120363161629617396", "120363161854316501", "120363165358271161", "120363168364416063", "120363169319669622", "120363169975121665", "120363171769248504", "120363189991282340", "120363196338530920", "120363199791777094", "120363236090549442", "120363242831731954", "120363260509199299", "120363283899919854", "120363291157960202", "120363294282893350", "120363310566841953", "12037188989", "120581458518255", "12065937224", "12088549831", "12089234440", "12092759604", "12104598191", "12107693415", "12144349391", "121616528871652", "12164180421", "12182030481", "12193011016", "12242691136", "12289349643", "122939395555337", "12312851586", "123463398334515", "12347552698", "124322341462059", "12485951624", "12512653471", "125177073500366", "12519578826", "12533985096", "12563428154", "12563481926", "12569600342", "12673489822", "12677187343", "12816587274", "12817444681", "129098345132130", "13035026845", "130369789644926", "13078281042", "13086550898", "13124682262", "13135550002", "13135550005", "13135550009", "13135550012", "13135550013", "13135550014", "13135550015", "13135550019", "13135550022", "13135550023", "13135550027", "13135550030", "13135550031", "13135550033", "13135550035", "13135550039", "13135550042", "13135550043", "13135550046", "13135550047", "13135550048", "13135550049", "13135550053", "13135550054", "13135550055", "13135550062", "13135550065", "13135550067", "13135550070", "13135550082", "13135550085", "13135550086", "13135550091", "13135550092", "13135550093", "13135550096", "13135550098", "13135550100", "13135550302", "13135550303", "13135550308", "13135550309", "13135550310", "13135550312", "13135550313", "13135550314", "13135550315", "13135550316", "13135550317", "13135550320", "13135550321", "13135550322", "13135550323", "13135550324", "13135550325", "13135550326", "13135550327", "13135550328", "13135550330", "13135550334", "13135550338", "13135559000", "13135559002", "13135559003", "13135559004", "13135559005", "13135559008", "13135559009", "13135559010", "13135559011", "13135559012", "13135559013", "13135559017", "13135559018", "13135559019", "131512150323454", "13179120539", "13179380030", "13187237549", "13188336425", "131988992307398", "13202417349", "132083330588794", "13232178658", "13234051120", "13234459871", "13234851291", "13236779632", "13237429682", "13238260610", "13346095713", "13367064708", "13373930303", "13465929760", "13472347237", "13472909155", "13473104333", "13474564797", "13475168420", "13475880967", "13475936676", "13478787368", "13479516322", "13526015327", "135459007143970", "137512169304289", "13852401837", "13858310873", "139225945108501", "139607945548011", "139998804332546", "14014191471", "14016489408", "14025048783", "14027821192", "14048864781", "14054665132", "14072702768", "14076006497", "14078215797", "14085991338", "14142101524", "14159910786", "14164479785", "141940297347204", "14195158467", "14232189459", "14235022702", "14242362362", "14242363754", "14244990541", "14256244006", "143181576454319", "14356776413", "14356776955", "14359905397", "14359905938", "143834445017308", "14438380391", "144792122048696", "14692785231", "147004080578653", "14736234139697", "14805990841", "14806561073", "14847577922", "150053339533511", "15028019221", "150302598627464", "15038471839", "15038640034", "15042332249", "15057175374", "15076217539", "15076975472", "15083061090", "150839452774560", "150925352136784", "15099617093", "15103146582", "15104844091", "15105924830", "15106214216", "15109340002", "15123879210", "15125224259", "15125713374", "15125716517", "15126553035", "15138765723", "15155235640", "15155259124", "15167789277", "15168005517", "15168799747", "15168886942", "15203295777", "15203555498", "152046388932662", "152488787329139", "15302068722", "15404237607", "15595540556", "15597596875", "15598199187", "15612890324", "15618035894", "15623286070", "15673433595", "15704602737", "15704977866", "15705409111", "15705947881", "15713298742", "15713349815", "15713480045", "15713559735", "15715446505", "15715646023", "15715917168", "15715980952", "15716062052", "15739197568", "16012070456", "16017304815", "16017809365", "16028162053", "16029186659", "16073820970", "16083350264", "16083597914", "16093461419", "16093505541", "16099336348", "16104573925", "16104835210", "16106046786", "16123546753", "16143169059", "16143683432", "16144467091", "16145003405", "16145921304", "16146225841", "16152891229", "16153625864", "16155107778", "16158818625", "161688489824333", "16175941188", "16193489516", "161954777829383", "16195989232", "162135283859693", "16235229187", "16262906591", "16263678595", "16263678786", "16263678865", "16264191192", "16266766304", "16269778212", "16302169283", "16304122840", "16308639426", "16319973442", "16462411419", "16463319077", "16467602090", "16468012779", "164939880726537", "16504166544", "16506848121", "16514436258", "16572062143", "16615243843", "166292694802535", "16782038340", "16784370740", "16789395667", "16822359596", "168229909594129", "168345789841472", "169548330315856", "17015809945", "17017387027", "170201081483270", "170265724104705", "17034241981", "17034473529", "17036534502", "17043887688", "170480405340393", "17064662606", "17066755548", "17072438970", "17075022524", "17076530463", "17133853574", "17142935893", "17147523683", "17149864683", "17162175691", "17173488600", "17183005254", "17185527136", "17192528628", "17192973293", "17207425163", "17244674519", "17272651415", "17315992272", "17323557893", "17328120994", "17329001069", "17345525321", "17408335494", "17408458768", "17423794330", "174869845135609", "175342325080150", "17547778147", "17622338037", "17654766272", "17655603474", "17694875554", "17694875893", "17706484589", "177159045906534", "17723629659", "17732708847", "17736696040", "17738778243", "178185543102533", "17862039442", "17862532958", "17863899629", "17864023107", "17865717549", "17865718190", "17865821967", "17866412763", "179001368785019", "18014276170", "18022898358", "18024518808", "18025880528", "18027250068", "18045888536", "18047188072", "18052538018", "18053218602", "18056377243", "18067510314", "18085096467", "18129021981", "18134709062", "18135326643", "18135512847", "18139660839", "18165228437", "18173535149", "18175659571", "18188148927", "18189341683", "1821301047437", "182201337524439", "182282774175819", "182390332911843", "182506246656145", "182605131567300", "18284752696", "18286800850", "18312041352", "18313053342", "18317078315", "18317721351", "18328059892", "184052434886731", "184245708431525", "18435802562", "18458017418", "185499872415779", "18562681088", "18565341559", "18565538852", "18566393015", "18569973113", "18588636236", "18602817442", "18603564643", "18605582168", "18624338328", "186302997745696", "18632531704", "18635129576", "18644836084", "18647474777", "186672415256810", "186942763302988", "18702815976", "18722227422", "18747783934027", "18769174892625", "190048343453815", "19014843965", "19033263809", "19036468314", "19048785221", "19082801531", "19082968935", "19096144641", "19096309946", "19099140470", "19108982297", "19139153988", "19143864069", "19144139560", "19152920437", "19159792153", "19162048632", "19167706237", "19185653253", "19198413910", "19198856271", "19199037779", "192388983173249", "19254368625", "19254779421", "192573649993951", "19317069676", "193273729683514", "19344140808", "19372305774", "194536433307865", "19494443119", "194944572596311", "19496190099", "195232134082660", "195245220360233", "19542343915", "19549552526", "19566170438", "19567321403", "19567893122", "19568214170", "19719008724", "19735203731", "197362521739402", "19785150566", "197873673183315", "19804012349", "19842441148", "19852369962", "198535098151027", "199252324159534", "199548693651693", "200467799892008", "202907442000051", "204582412144684", "204848750411844", "206150142304493", "206377708470397", "206493689376827", "206781385048068", "208018385956938", "208632583032844", "209925384999102", "215014870888498", "215306978984178", "215319880687714", "215410024661201", "217218222682204", "217239731085336", "217617637859387", "218545384329241", "221667792027769", "221697705824396", "221856770617483", "224523945291787", "225391562211456", "227414475075628", "228548329611468", "22995271712920", "230476769964063", "231623526224069", "231971435364603", "232465323028575", "237915703664686", "239603525148848", "240432336367649", "241282874126532", "241587833577501", "24378335039589", "243958638731459", "245491942039636", "246827693678606", "253772672573507", "253944488009965", "255619357483108", "256993730236621", "257384723259421", "259356130029573", "259652516356294", "260972379022", "262452885336154", "263178684469434", "264020498092255", "266833718431948", "26710653329462", "268246779445272", "268951053439040", "269307569266755", "269586725388507", "271734125146250", "274349860909238", "275866017865913", "276029226676359", "277167376183308", "279774371000453", "279783011262693", "28531702653037", "29764358254642", "30464505073828", "32233796698326", "33785650921", "34836714635416", "35639856766996", "39844579401782", "4033209221175", "40832505757728", "40978350141447", "41042975949001", "4264919351425", "43697081192639", "43890354696378", "44418736336924", "447974045725", "45908939030566", "4915257482547", "491747383347", "491747383354", "491747386191", "491748589154", "491748599935", "491748600734", "491748672588", "49826133741698", "50350170058985", "5213311857022", "52600531632270", "54692432318606", "56410419257463", "56616359583845", "58613636853903", "6287745672611", "64360370192402", "65476977782923", "69999628665014", "71983886766200", "72928746037443", "74144305684519", "74917131374625", "75273798172747", "76742727315704", "77554526470395", "79401278517497", "8001775771719", "81367430271", "82506523119859", "85869633482900", "90452212613311", "905338541080", "91057954017452", "923130999544", "923323013167", "9534928064606", "96529440321619", "status", "Svetlana Chernoff", "LTC Whalen", "Capt Don Wayne", "Karen Tate", "Hank", "Brian Reynolds", "Goldie Kahn", "Vladamir Stravinsky", "6️⃣ Wealth Builders Club", "Joey", "Mary Garcia", "Abe Rudder", "Russ Philby", "Voice Mail", "Otto", "Toks", "Meta", "leaura", "faxinezidohne", "patrickh34", "yauyauyauhen", "nathanmorris", "italianmatters", "robm435", "loususi", "lanaire2023", "pet_the_bunny", "madmax_mgm", "reidback", "rennymorales", "yjr_fit.inba", "the_real_flockfam", "humans_of_data", "airwicksol", "homan.jason", "nadhir_chiu_oficial", "brandonmcclainl", "geezdagawd", "superflysugar2024", "curia__", "psychicadvisor345", "lemieuxbrands", "james_macray_"], "Num_of_PII": 680, "source_columns": ["wa_contacts.jid", "wa_contacts.display_name", "wa_contacts.nickname", "wa_address_book.jid", "wa_address_book.display_name", "wa_address_book.nickname", "wa_biz_profiles.jid", "wa_biz_profiles.custom_url", "wa_bot_profiles.creator_name"], "Raw_rows_first_100": [["100674318659756@lid"], ["103023648964858@lid"], ["103929853546678@lid"], ["103981426733067@lid"], ["104569854005296@lid"], ["106940592033823@lid"], ["107967072444508@lid"], ["108272048668771@lid"], ["109766730866918@lid"], ["110638642757646@lid"], ["112103243395097@lid"], ["11356128460981@lid"], ["115504622653502@lid"], ["115728095129747@lid"], ["116032903606475@lid"], ["12014946184@s.whatsapp.net"], ["12025692832@s.whatsapp.net"], ["12028177932@s.whatsapp.net"], ["12028275725@s.whatsapp.net"], ["120363098389106519@g.us"], ["120363144038483540@newsletter"], ["120363159552753674@newsletter"], ["120363160757210405@newsletter"], ["120363161629617396@newsletter"], ["120363161854316501@newsletter"], ["120363165358271161@newsletter"], ["120363168364416063@newsletter"], ["120363169319669622@newsletter"], ["120363169975121665@newsletter"], ["120363171769248504@newsletter"], ["120363189991282340@newsletter"], ["120363196338530920@newsletter"], ["120363199791777094@newsletter"], ["120363236090549442@newsletter"], ["120363242831731954@newsletter"], ["120363260509199299@newsletter"], ["120363283899919854@newsletter"], ["120363291157960202@newsletter"], ["120363294282893350@newsletter"], ["120363310566841953@newsletter"], ["12037188989@s.whatsapp.net"], ["120581458518255@lid"], ["12065937224@s.whatsapp.net"], ["12088549831@s.whatsapp.net"], ["12089234440@s.whatsapp.net"], ["12092759604@s.whatsapp.net"], ["12104598191@s.whatsapp.net"], ["12107693415@s.whatsapp.net"], ["12144349391@s.whatsapp.net"], ["121616528871652@lid"], ["12164180421@s.whatsapp.net"], ["12182030481@s.whatsapp.net"], ["12193011016@s.whatsapp.net"], ["12242691136@s.whatsapp.net"], ["12289349643@s.whatsapp.net"], ["122939395555337@lid"], ["12312851586@s.whatsapp.net"], ["123463398334515@lid"], ["12347552698@s.whatsapp.net"], ["124322341462059@lid"], ["12485951624@s.whatsapp.net"], ["12512653471@s.whatsapp.net"], ["125177073500366@lid"], ["12519578826@s.whatsapp.net"], ["12533985096@s.whatsapp.net"], ["12563428154@s.whatsapp.net"], ["12563481926@s.whatsapp.net"], ["12569600342@s.whatsapp.net"], ["12673489822@s.whatsapp.net"], ["12677187343@s.whatsapp.net"], ["12816587274@s.whatsapp.net"], ["12817444681@s.whatsapp.net"], ["129098345132130@lid"], ["13035026845@s.whatsapp.net"], ["130369789644926@lid"], ["13078281042@s.whatsapp.net"], ["13086550898@s.whatsapp.net"], ["13124682262@s.whatsapp.net"], ["13135550002@s.whatsapp.net"], ["13135550005@s.whatsapp.net"], ["13135550009@s.whatsapp.net"], ["13135550012@s.whatsapp.net"], ["13135550013@s.whatsapp.net"], ["13135550014@s.whatsapp.net"], ["13135550015@s.whatsapp.net"], ["13135550019@s.whatsapp.net"], ["13135550022@s.whatsapp.net"], ["13135550023@s.whatsapp.net"], ["13135550027@s.whatsapp.net"], ["13135550030@s.whatsapp.net"], ["13135550031@s.whatsapp.net"], ["13135550033@s.whatsapp.net"], ["13135550035@s.whatsapp.net"], ["13135550039@s.whatsapp.net"], ["13135550042@s.whatsapp.net"], ["13135550043@s.whatsapp.net"], ["13135550046@s.whatsapp.net"], ["13135550047@s.whatsapp.net"], ["13135550048@s.whatsapp.net"], ["13135550049@s.whatsapp.net"]], "Total_raw_rows": 842, "Exploration_sql": "SELECT jid FROM wa_contacts WHERE jid REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b' \nUNION ALL \nSELECT display_name FROM wa_contacts WHERE display_name REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b' \nUNION ALL \nSELECT nickname FROM wa_contacts WHERE nickname REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b' \nUNION ALL \nSELECT jid FROM wa_address_book WHERE jid REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b' \nUNION ALL \nSELECT display_name FROM wa_address_book WHERE display_name REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b' \nUNION ALL \nSELECT nickname FROM wa_address_book WHERE nickname REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b' \nUNION ALL \nSELECT jid FROM wa_biz_profiles WHERE jid REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b' \nUNION ALL \nSELECT custom_url FROM wa_biz_profiles WHERE custom_url REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b' \nUNION ALL \nSELECT creator_name FROM wa_bot_profiles WHERE creator_name REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b';", "Extraction_sql": "SELECT jid FROM wa_contacts WHERE jid REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT display_name FROM wa_contacts WHERE display_name REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT nickname FROM wa_contacts WHERE nickname REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT jid FROM wa_address_book WHERE jid REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT display_name FROM wa_address_book WHERE display_name REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT nickname FROM wa_address_book WHERE nickname REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT jid FROM wa_biz_profiles WHERE jid REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT custom_url FROM wa_biz_profiles WHERE custom_url REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b'\nUNION ALL\nSELECT creator_name FROM wa_bot_profiles WHERE creator_name REGEXP '\\b[a-zA-Z][a-zA-Z0-9._-]{2,51}\\b';", "PII_Prompt": "a username is a short textual identifier chosen by a user to represent their account or public handle within an application or service it is stored as plain text contains no whitespace does not include a domain component and is intended for human recognition rather than internal system uniqueness"}
{"db_path": "selectedDBs\\A1_wa.db", "PII_type": "PERSON_NAME", "PII": ["Svetlana Chernoff", "LTC Whalen", "Capt Don Wayne", "Karen Tate", "Brian Reynolds", "Goldie Kahn", "Vladamir Stravinsky", "Mary Garcia", "Abe Rudder", "Russ Philby"], "Num_of_PII": 10, "source_columns": ["wa_contacts.display_name", "wa_contacts.given_name", "wa_contacts.family_name", "wa_contacts.nickname", "wa_address_book.display_name", "wa_address_book.given_name", "wa_address_book.family_name", "wa_address_book.nickname", "wa_biz_profiles.business_description", "wa_biz_profiles.location_name"], "Raw_rows_first_100": [["Svetlana Chernoff"], ["LTC Whalen"], ["Capt Don Wayne"], ["Karen Tate"], ["Hank"], ["Brian Reynolds"], ["Goldie Kahn"], ["Vladamir Stravinsky"], ["6️⃣ Wealth Builders Club"], ["Joey"], ["Mary Garcia"], ["Abe Rudder"], ["Russ Philby"], ["Voice Mail"], ["Otto"], ["Toks"], ["Bo"], ["Svetlana"], ["Don"], ["Karen"], ["Hank"], ["Brian"], ["Goldie"], ["Vladamir"], ["Joey"], ["Mary"], ["Abe"], ["Russ"], ["Voice Mail"], ["Otto"], ["Toks"], ["Bo"], ["Chernoff"], ["Whalen"], ["Wayne"], ["Tate"], ["Reynolds"], ["Kahn"], ["Stravinsky"], ["Garcia"], ["Rudder"], ["Philby"], ["Russ Philby"], ["Capt Don Wayne"], ["LTC Whalen"], ["Mary Garcia"], ["Joey"], ["Hank"], ["Goldie Kahn"], ["Abe Rudder"], ["Svetlana Chernoff"], ["Karen Tate"], ["Voice Mail"], ["Brian Reynolds"], ["Vladamir Stravinsky"], ["Otto"], ["Toks"], ["Bo"], ["Russ"], ["Don"], ["Mary"], ["Joey"], ["Hank"], ["Goldie"], ["Abe"], ["Svetlana"], ["Karen"], ["Voice Mail"], ["Brian"], ["Vladamir"], ["Otto"], ["Toks"], ["Bo"], ["Philby"], ["Wayne"], ["Whalen"], ["Garcia"], ["Kahn"], ["Rudder"], ["Chernoff"], ["Tate"], ["Reynolds"], ["Stravinsky"], ["Dispatch and Sales"], ["Dispatch and Sales"]], "Total_raw_rows": 85, "Exploration_sql": "SELECT display_name FROM wa_contacts WHERE display_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}' \nUNION ALL \nSELECT given_name FROM wa_contacts WHERE given_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}' \nUNION ALL \nSELECT family_name FROM wa_contacts WHERE family_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}' \nUNION ALL \nSELECT nickname FROM wa_contacts WHERE nickname REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}' \nUNION ALL \nSELECT display_name FROM wa_address_book WHERE display_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}' \nUNION ALL \nSELECT given_name FROM wa_address_book WHERE given_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}' \nUNION ALL \nSELECT family_name FROM wa_address_book WHERE family_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}' \nUNION ALL \nSELECT nickname FROM wa_address_book WHERE nickname REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}' \nUNION ALL \nSELECT business_description FROM wa_biz_profiles WHERE business_description REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}' \nUNION ALL \nSELECT location_name FROM wa_biz_profiles WHERE location_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}';", "Extraction_sql": "SELECT display_name FROM wa_contacts WHERE display_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}'\nUNION ALL\nSELECT given_name FROM wa_contacts WHERE given_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}'\nUNION ALL\nSELECT family_name FROM wa_contacts WHERE family_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}'\nUNION ALL\nSELECT nickname FROM wa_contacts WHERE nickname REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}'\nUNION ALL\nSELECT display_name FROM wa_address_book WHERE display_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}'\nUNION ALL\nSELECT given_name FROM wa_address_book WHERE given_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}'\nUNION ALL\nSELECT family_name FROM wa_address_book WHERE family_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}'\nUNION ALL\nSELECT nickname FROM wa_address_book WHERE nickname REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}'\nUNION ALL\nSELECT business_description FROM wa_biz_profiles WHERE business_description REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}'\nUNION ALL\nSELECT location_name FROM wa_biz_profiles WHERE location_name REGEXP '[A-Za-z][A-Za-z\\s\\.\\-]{1,50}';", "PII_Prompt": "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.)"}
{"db_path": "selectedDBs\\A1_wa.db", "PII_type": "POSTAL_ADDRESS", "PII": [], "Num_of_PII": 0, "source_columns": [], "Raw_rows_first_100": [], "Total_raw_rows": 0, "Exploration_sql": "SELECT display_name FROM wa_contacts WHERE display_name REGEXP '(?i)\\b(?:p\\.?\\s*o\\.?\\s*box|post\\s+office\\s+box|ave\\.?|avenue|st\\.?|street|rd\\.?|road|blvd\\.?|boulevard|dr\\.?|drive|ln\\.?|lane|ct\\.?|court|pl\\.?|place|way|pkwy\\.?|parkway|cir\\.?|circle|ter\\.?|terrace|hwy\\.?|highway|trl\\.?|trail|sq\\.?|square|pike|loop|run|walk|path|byp\\.?|bypass|(?:n|s|e|w|ne|nw|se|sw)\\b)\\b' \nUNION ALL \nSELECT company FROM wa_contacts WHERE company REGEXP '(?i)\\b(?:p\\.?\\s*o\\.?\\s*box|post\\s+office\\s+box|ave\\.?|avenue|st\\.?|street|rd\\.?|road|blvd\\.?|boulevard|dr\\.?|drive|ln\\.?|lane|ct\\.?|court|pl\\.?|place|way|pkwy\\.?|parkway|cir\\.?|circle|ter\\.?|terrace|hwy\\.?|highway|trl\\.?|trail|sq\\.?|square|pike|loop|run|walk|path|byp\\.?|bypass|(?:n|s|e|w|ne|nw|se|sw)\\b)\\b' \nUNION ALL \nSELECT display_name FROM wa_address_book WHERE display_name REGEXP '(?i)\\b(?:p\\.?\\s*o\\.?\\s*box|post\\s+office\\s+box|ave\\.?|avenue|st\\.?|street|rd\\.?|road|blvd\\.?|boulevard|dr\\.?|drive|ln\\.?|lane|ct\\.?|court|pl\\.?|place|way|pkwy\\.?|parkway|cir\\.?|circle|ter\\.?|terrace|hwy\\.?|highway|trl\\.?|trail|sq\\.?|square|pike|loop|run|walk|path|byp\\.?|bypass|(?:n|s|e|w|ne|nw|se|sw)\\b)\\b' \nUNION ALL \nSELECT company FROM wa_address_book WHERE company REGEXP '(?i)\\b(?:p\\.?\\s*o\\.?\\s*box|post\\s+office\\s+box|ave\\.?|avenue|st\\.?|street|rd\\.?|road|blvd\\.?|boulevard|dr\\.?|drive|ln\\.?|lane|ct\\.?|court|pl\\.?|place|way|pkwy\\.?|parkway|cir\\.?|circle|ter\\.?|terrace|hwy\\.?|highway|trl\\.?|trail|sq\\.?|square|pike|loop|run|walk|path|byp\\.?|bypass|(?:n|s|e|w|ne|nw|se|sw)\\b)\\b';", "Extraction_sql": null, "PII_Prompt": "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')"}
Reference in New Issue View Git Blame Copy Permalink