Dom/anonymisation
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: Optimize EPISODE false positives - filter trackare filename episodes

Browse Source 
- Modified detectors/hospital_filter.py:
  * Updated is_episode_in_filename() to only filter trackare documents
  * Pattern: trackare-XXXXXXXX-YYYYYYYY where YYYYYYYY is episode number
  * Prevents filtering legitimate episodes in CRH/CRO documents

- Modified anonymizer_core_refactored_onnx.py:
  * Filter page=-1 entries (global propagation) from audit file
  * These are internal replacement tokens, not real detections

- Modified evaluation/quality_evaluator.py:
  * Fixed load_annotations() to use ground_truth_dir instead of pdf_path.parent
  * Added support for 'pages' format from auto-annotation script
  * Converts 'pages' format to 'annotations' format automatically

- Updated test dataset annotations with hospital filter applied

Results:
- EPISODE: Precision 100% (was 14.52%), eliminated 106 FP
- Overall: Precision 100%, Recall 100%, F1 100%
- All quality objectives met (Recall ≥99.5%, Precision ≥97%, F1 ≥98%)
This commit is contained in:
Domi31tls
2026-03-02 15:33:29 +01:00
parent 883f14ab79
commit ee34042179
97 changed files with 2140 additions and 9878 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
tests/ground_truth/annotations/001_simple_unknown_BACTERIO_23018396.json
View File

@@ -10,9 +10,6 @@
"ETABLISSEMENT": [
"Centre Hospitalier de la Côte Basque"
],
"TEL": [
"0559443674"
],
"NOM": [
"JAOUEN Anne-Christine",
"MENARD-DEROURE Fanny",

3
tests/ground_truth/annotations/002_simple_unknown_bacterio_476_23159413.json
View File

@@ -10,9 +10,6 @@
"ETABLISSEMENT": [
"Centre Hospitalier de la Côte Basque"
],
"TEL": [
"0559443674"
],
"NOM": [
"JAOUEN Anne-Christine",
"MENARD-DEROURE Fanny",

38
tests/ground_truth/annotations/005_simple_compte_rendu_CRH_23155836.json
View File

@@ -7,23 +7,6 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"102 RUE MARIE CURIE"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"40390 ST MARTIN DE SEIGNANX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.23",
"05.59.44.37.25",
"05.59.44.37.22",
"05.59.44.37.29"
],
"ETABLISSEMENT": [
"Pôle Spécialités Médicales",
"Service de Gastro-Entérologie - Oncologie Digestive"
@@ -36,6 +19,12 @@
"BRUGEL",
"GUILNGAR"
],
"ADRESSE": [
"102 RUE MARIE CURIE"
],
"CODE_POSTAL": [
"40390 ST MARTIN DE SEIGNANX"
],
"DATE_NAISSANCE": [
"née le 27/04/1959"
],
@@ -65,21 +54,6 @@
{
"page_number": 1,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.23",
"05.59.44.37.25",
"05.59.44.37.22",
"05.59.44.37.29"
],
"ETABLISSEMENT": [
"Pôle Spécialités Médicales",
"Service de Gastro-Entérologie - Oncologie Digestive"

20
tests/ground_truth/annotations/008_simple_trackare_trackare-14004105-23202435_14004105_23202435.json
View File

@@ -7,18 +7,6 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"22 LOT MENDI ALDE Ville de résidence",
"4, AVENUE DE TRÉVILLE ",
"22 LOT MENDI ALDE\tVille de résidence"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"Code Postal: 64130",
"64130 MAULEON-LICHARRE"
],
"IPP": [
"14004105"
],
@@ -28,6 +16,14 @@
"VILLE": [
"CHERAUTE"
],
"CODE_POSTAL": [
"Code Postal: 64130",
"64130 MAULEON-LICHARRE"
],
"ADRESSE": [
"22 LOT MENDI ALDE Ville de résidence",
"22 LOT MENDI ALDE\tVille de résidence"
],
"NOM": [
"Romain DIDAILLER",
"François GARNIER"

9
tests/ground_truth/annotations/010_simple_anapath_ANAPATH_23217289.json
View File

@@ -16,16 +16,15 @@
"DIDAILLER Romain",
"Lewis GRECOURT"
],
"ADRESSE": [
"13 Av. de l'Interne Jacques Loeb",
"14 allée de Bordenave ",
"14 allée de bordenave "
],
"CODE_POSTAL": [
"64100 BAYONNE",
"64240 MACAYE",
"64990 SAINT PIERRE"
],
"ADRESSE": [
"14 allée de Bordenave ",
"14 allée de bordenave "
],
"TEL": [
"05 24 33 03 91"
]

24
tests/ground_truth/annotations/012_moyen_compte_rendu_CRH_692_23200418.json
View File

@@ -7,28 +7,12 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13 avenue de l",
"4, ALLÉE BORDENAVE"
],
"CODE_POSTAL": [
"64109 BAYONNE Cedex",
"64990 ST PIERRE"
],
"ETABLISSEMENT": [
"Pôle de Chirurgie - Anesthésie - Bloc Opératoire",
"Unité Urologie"
],
"TEL": [
"05.59.44.38.44",
"05.59.4 4.35.23",
"05.59.44.35.05",
"05.59.44.35.03",
"05.59.44.44.94",
"05.59.44.43.42",
"05.59.44.35.02",
"05.59.44.35.09",
"05.59.44.32.01"
"05.59.4 4.35.23"
],
"NOM": [
"Romain DIDAILLER",
@@ -46,6 +30,12 @@
"Florence MAZERES",
"Caroline RIVERA",
"Bruno CORDON"
],
"ADRESSE": [
"4, ALLÉE BORDENAVE"
],
"CODE_POSTAL": [
"64990 ST PIERRE"
]
}
}

27
tests/ground_truth/annotations/013_moyen_compte_rendu_363_23085243_CRO.json
View File

@@ -7,30 +7,10 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13 avenue de l",
"4 RUE DE BELFORT",
"6, CHEMIN DE LA MAROUETTE"
],
"CODE_POSTAL": [
"64109 BAYONNE Cedex",
"64100 BAYONNE"
],
"ETABLISSEMENT": [
"Pôle de Chirurgie - Anesthésie - Bloc Opératoire",
"Unité Urologie"
],
"TEL": [
"05.59.44.38.44",
"05.59.44.35.23",
"05.59.44.35.05",
"05.59.44.35.03",
"05.59.44.44.94",
"05.59.44.43.42",
"05.59.44.35.02",
"05.59.44.35.09",
"05.59.44.32.01"
],
"NOM": [
"Romain DIDAILLER",
"Laura ETCHECHOURY",
@@ -48,6 +28,13 @@
"Caroline RIVERA",
"Bruno CORDON"
],
"ADRESSE": [
"4 RUE DE BELFORT",
"6, CHEMIN DE LA MAROUETTE"
],
"CODE_POSTAL": [
"64100 BAYONNE"
],
"DATE_NAISSANCE": [
"Né le 28/03/1942"
]

70
tests/ground_truth/annotations/016_moyen_compte_rendu_CRH_23149905.json
View File

@@ -7,29 +7,6 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"2 AVENUE PIERRE LARRAMENDY"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.35.69",
"05.59.44.35.30",
"05.59.44.35.06",
"05.59.44.39.24",
"05.59.44.37.07",
"05.59.44.37.33",
"05.59.44.31.39",
"05.59.44.37.35",
"05.59.44.37.46",
"05.59.44.37.32",
"05.59.44.37.39"
],
"ETABLISSEMENT": [
"Pôle de Médecine Interne",
"Service de Maladies Infectieuses",
@@ -48,6 +25,9 @@
"Heidi WILLE IRC",
"Claire CASTEL"
],
"ADRESSE": [
"2 AVENUE PIERRE LARRAMENDY"
],
"RPPS": [
"10101718855",
"10101489531",
@@ -73,28 +53,6 @@
{
"page_number": 1,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.35.69",
"05.59.44.35.30",
"05.59.44.35.06",
"05.59.44.39.24",
"05.59.44.37.07",
"05.59.44.37.33",
"05.59.44.31.39",
"05.59.44.37.35",
"05.59.44.37.46",
"05.59.44.37.32",
"05.59.44.37.39"
],
"ETABLISSEMENT": [
"Pôle de Médecine Interne",
"Service de Maladies Infectieuses",
@@ -131,28 +89,6 @@
{
"page_number": 2,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.35.69",
"05.59.44.35.30",
"05.59.44.35.06",
"05.59.44.39.24",
"05.59.44.37.07",
"05.59.44.37.33",
"05.59.44.31.39",
"05.59.44.37.35",
"05.59.44.37.46",
"05.59.44.37.32",
"05.59.44.37.39"
],
"ETABLISSEMENT": [
"Pôle de Médecine Interne",
"Service de Maladies Infectieuses",

71
tests/ground_truth/annotations/018_moyen_compte_rendu_CRH_23042753.json
View File

@@ -7,23 +7,6 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"3297 QUARTIER AUZO TTIPI"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"64430 ST ETIENNE DE BAIGORRY"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.23",
"05.59.44.37.25",
"05.59.44.37.22",
"05.59.44.37.29"
],
"ETABLISSEMENT": [
"Pôle Spécialités Médicales",
"Service de Gastro-Entérologie - Oncologie Digestive"
@@ -37,6 +20,12 @@
"NIVET",
"PUJOS"
],
"ADRESSE": [
"3297 QUARTIER AUZO TTIPI"
],
"CODE_POSTAL": [
"64430 ST ETIENNE DE BAIGORRY"
],
"DATE_NAISSANCE": [
"née le 23/02/1980"
],
@@ -65,20 +54,6 @@
{
"page_number": 1,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.25",
"05.59.44.37.22",
"05.59.44.37.29"
],
"ETABLISSEMENT": [
"Pôle Spécialités Médicales",
"Service de Gastro-Entérologie - Oncologie Digestive"
@@ -118,22 +93,6 @@
{
"page_number": 2,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"64430 ST ETIENNE DE BAIGORRY"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.23",
"05.59.44.37.25",
"05.59.44.37.22",
"05.59.44.37.29"
],
"ETABLISSEMENT": [
"Pôle Spécialités Médicales",
"Service de Gastro-Entérologie - Oncologie Digestive"
@@ -146,6 +105,9 @@
"NIVET",
"PUJOS"
],
"CODE_POSTAL": [
"64430 ST ETIENNE DE BAIGORRY"
],
"DATE_NAISSANCE": [
"née le 23/02/1980"
],
@@ -173,21 +135,6 @@
{
"page_number": 3,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.23",
"05.59.44.37.25",
"05.59.44.37.22",
"05.59.44.37.29"
],
"ETABLISSEMENT": [
"Pôle Spécialités Médicales",
"Service de Gastro-Entérologie - Oncologie Digestive"

20
tests/ground_truth/annotations/019_moyen_compte_rendu_CRO_332_23049003.json
View File

@@ -34,16 +34,6 @@
"ADRESSE": [
"1286 CHEMIN DE GAINEKO BORDA"
],
"TEL": [
"05.59.44.33.20",
"05.59.44.35.43",
"05.59.44.35.47",
"05.59.44.43.58",
"05.59.44.35.49",
"05.59.44.43.44",
"05.59.44.35.42",
"05.59.44.35.45"
],
"DATE_NAISSANCE": [
"né le 26/08/1947"
],
@@ -76,16 +66,6 @@
"AGE": [
"Patient de 75 ans"
],
"TEL": [
"05.59.44.33.20",
"05.59.44.35.43",
"05.59.44.35.47",
"05.59.44.43.58",
"05.59.44.35.49",
"05.59.44.43.44",
"05.59.44.35.42",
"05.59.44.35.45"
],
"EMAIL": [
"secr.neurochir@ch-cotebasque.fr"
]

3
tests/ground_truth/annotations/022_moyen_compte_rendu_cro2_516_23187028.json
View File

@@ -11,9 +11,6 @@
"Centre Hospitalier de la Côte Basque",
"Service Demandeur"
],
"TEL": [
"05.59.44.35.35"
],
"NOM": [
"Samuel KASPARIAN"
]

174
tests/ground_truth/annotations/023_complexe_compte_rendu_CRH_23102610.json
View File

@@ -7,25 +7,6 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"24 AVENUE DE LA BAIE DE TXIGUNDI"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"64700 HENDAYE"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.33",
"05.59.44.37.42",
"05.59.44.37.32",
"05.59.44.38.62",
"05.59.44.37.74",
"05.33.78.81.89"
],
"ETABLISSEMENT": [
"Pôle Médecine Interne",
"Service Dyslipidémie",
@@ -57,6 +38,12 @@
"Loiseau",
"Moldovane"
],
"ADRESSE": [
"24 AVENUE DE LA BAIE DE TXIGUNDI"
],
"CODE_POSTAL": [
"64700 HENDAYE"
],
"DATE_NAISSANCE": [
"né le 30/07/1950"
],
@@ -74,23 +61,6 @@
{
"page_number": 1,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.33",
"05.59.44.37.42",
"05.59.44.37.32",
"05.59.44.38.62",
"05.59.44.37.74",
"05.33.78.81.89"
],
"ETABLISSEMENT": [
"Pôle Médecine Interne",
"Service MV",
@@ -134,23 +104,6 @@
{
"page_number": 2,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.33",
"05.59.44.37.42",
"05.59.44.37.32",
"05.59.44.38.62",
"05.59.44.37.74",
"05.33.78.81.89"
],
"ETABLISSEMENT": [
"Pôle Médecine Interne",
"Service Docteur MAURY Elisa",
@@ -194,25 +147,6 @@
{
"page_number": 3,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"57 BOULEVARD GENERAL LECLERC"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"64700 HENDAYE"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.33",
"05.59.44.37.42",
"05.59.44.37.32",
"05.59.44.38.62",
"05.59.44.37.74",
"05.33.78.81.89"
],
"ETABLISSEMENT": [
"Pôle Médecine Interne",
"Service Dyslipidémie",
@@ -245,6 +179,12 @@
"Loiseau",
"Moldovane"
],
"ADRESSE": [
"57 BOULEVARD GENERAL LECLERC"
],
"CODE_POSTAL": [
"64700 HENDAYE"
],
"DATE_NAISSANCE": [
"né le 30/07/1950"
],
@@ -262,23 +202,6 @@
{
"page_number": 4,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.33",
"05.59.44.37.42",
"05.59.44.37.32",
"05.59.44.38.62",
"05.59.44.37.74",
"05.33.78.81.89"
],
"ETABLISSEMENT": [
"Pôle Médecine Interne",
"Service MV",
@@ -322,23 +245,6 @@
{
"page_number": 5,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.33",
"05.59.44.37.42",
"05.59.44.37.32",
"05.59.44.38.62",
"05.59.44.37.74",
"05.33.78.81.89"
],
"ETABLISSEMENT": [
"Pôle Médecine Interne",
"Service Docteur MAURY Elisa",
@@ -382,25 +288,6 @@
{
"page_number": 6,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"1 PLACE AMELIE RABA LEON"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"33076 BORDEAUX CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.33",
"05.59.44.37.42",
"05.59.44.37.32",
"05.59.44.38.62",
"05.59.44.37.74",
"05.33.78.81.89"
],
"ETABLISSEMENT": [
"Pôle Médecine Interne",
"Service Dyslipidémie",
@@ -432,6 +319,9 @@
"Loiseau",
"Moldovane"
],
"ADRESSE": [
"1 PLACE AMELIE RABA LEON"
],
"DATE_NAISSANCE": [
"né le 30/07/1950"
],
@@ -449,23 +339,6 @@
{
"page_number": 7,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.33",
"05.59.44.37.42",
"05.59.44.37.32",
"05.59.44.38.62",
"05.59.44.37.74",
"05.33.78.81.89"
],
"ETABLISSEMENT": [
"Pôle Médecine Interne",
"Service MV",
@@ -509,23 +382,6 @@
{
"page_number": 8,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX"
],
"TEL": [
"05 59 44 35 35",
"05 59 63 35 88",
"05.59.44.37.33",
"05.59.44.37.42",
"05.59.44.37.32",
"05.59.44.38.62",
"05.59.44.37.74",
"05.33.78.81.89"
],
"ETABLISSEMENT": [
"Pôle Médecine Interne",
"Service Docteur MAURY Elisa",

22
tests/ground_truth/annotations/024_complexe_trackare_trackare-17001141-23066188_17001141_23066188.json
View File

@@ -7,26 +7,20 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"33 RUE JEAN FOURCADE Ville de résidence",
"39 rue Bernard de Coral ",
"33 RUE JEAN FOURCADE\tVille de résidence"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"Code Postal: 64122",
"64122 URRUGNE"
],
"IPP": [
"17001141"
],
"DATE_NAISSANCE": [
"Date de naissance: 15/01/2017"
],
"VILLE": [
"BAYONNE CEDEX"
"CODE_POSTAL": [
"Code Postal: 64122",
"64122 URRUGNE"
],
"ADRESSE": [
"33 RUE JEAN FOURCADE Ville de résidence",
"39 rue Bernard de Coral ",
"33 RUE JEAN FOURCADE\tVille de résidence"
],
"NOM": [
"Céline BELLEAU",

21
tests/ground_truth/annotations/025_complexe_trackare_trackare-02016820-23095226_02016820_23095226.json
View File

@@ -7,18 +7,6 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"7 RUE DES PADOUANS Ville de résidence",
"12 rue de l'industrie ",
"7 RUE DES PADOUANS\tVille de résidence"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"Code Postal: 64100",
"64600 ANGLET"
],
"IPP": [
"02016820"
],
@@ -28,6 +16,15 @@
"VILLE": [
"OLORON STE MARIE"
],
"CODE_POSTAL": [
"Code Postal: 64100",
"64600 ANGLET"
],
"ADRESSE": [
"7 RUE DES PADOUANS Ville de résidence",
"12 rue de l'industrie ",
"7 RUE DES PADOUANS\tVille de résidence"
],
"NOM": [
"Laurence MASSE",
"Gilles DELMAS"

22
tests/ground_truth/annotations/026_complexe_trackare_trackare-15000536-23074384_15000536_23074384.json
View File

@@ -7,26 +7,20 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"1 RUE JOSEPH ST ANDRÉ Ville de résidence",
"4 RUE PONTRIQUE ",
"1 RUE JOSEPH ST ANDRÉ\tVille de résidence"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"Code Postal: 64340",
"64100 BAYONNE"
],
"IPP": [
"15000536"
],
"DATE_NAISSANCE": [
"Date de naissance: 08/01/2015"
],
"VILLE": [
"BAYONNE CEDEX"
"CODE_POSTAL": [
"Code Postal: 64340",
"64100 BAYONNE"
],
"ADRESSE": [
"1 RUE JOSEPH ST ANDRÉ Ville de résidence",
"4 RUE PONTRIQUE ",
"1 RUE JOSEPH ST ANDRÉ\tVille de résidence"
],
"NOM": [
"Marie DUBREL",

21
tests/ground_truth/annotations/027_complexe_trackare_trackare-10027557-23183041_10027557_23183041.json
View File

@@ -7,18 +7,6 @@
{
"page_number": 0,
"pii": {
"ADRESSE": [
"13, Avenue de l'Interne J",
"LOEB BP 8",
"4 RUE DU PETIT NANOT Ville de résidence",
"1, PLACE PEREIRE ",
"4 RUE DU PETIT NANOT\tVille de résidence"
],
"CODE_POSTAL": [
"64109 BAYONNE CEDEX",
"Code Postal: 64340",
"64100 BAYONNE"
],
"IPP": [
"10027557"
],
@@ -28,6 +16,15 @@
"VILLE": [
"PARIS"
],
"CODE_POSTAL": [
"Code Postal: 64340",
"64100 BAYONNE"
],
"ADRESSE": [
"4 RUE DU PETIT NANOT Ville de résidence",
"1, PLACE PEREIRE ",
"4 RUE DU PETIT NANOT\tVille de résidence"
],
"NOM": [
"Marie LACLAU-LACROUTS",
"Georges PEPIN"

12
tests/ground_truth/annotations/dataset_statistics.json
View File

@@ -1,23 +1,23 @@
{
"total_documents": 25,
"total_pages": 133,
"total_pii": 1167,
"total_pii": 907,
"by_type": {
"ETABLISSEMENT": 83,
"TEL": 193,
"NOM": 507,
"IPP": 25,
"ADRESSE": 79,
"CODE_POSTAL": 50,
"ADRESSE": 29,
"CODE_POSTAL": 24,
"DATE_NAISSANCE": 114,
"EMAIL": 62,
"RPPS": 21,
"EPISODE": 18,
"VILLE": 5,
"VILLE": 3,
"TEL": 11,
"AGE": 5,
"NIR": 2,
"DOSSIER": 3
},
"avg_pii_per_doc": 46.7,
"avg_pii_per_doc": 36.3,
"avg_pages_per_doc": 5.3
}