chore(bench): résultats V2 et rapports de benchmarking

Snapshot des 18 JSONs produits par le pipeline V2 (Qwen2.5-VL-3B +
checkboxes densité + validation ATIH), utiles au collaborateur comme
référence de ce que la chaîne actuelle produit.

Rapports :
- bench_v2_report.md       : comparaison V2 vs legacy docTR+VLM
                             (couverture, divergences, régressions
                             notables sur codage_reco et praticien).
- validation_report.md     : résumé de la validation ATIH sur les 18
                             JSONs (131/149 → 140/149 codes valides
                             après fix suffixes `*` et `+N`, 0
                             incohérence GHM↔GHS, 8 suggestions de
                             correction OCR).

Script de comparaison :
- bench_v11_vs_legacy.py   : tableau d'accord champ par champ entre
                             un run du pipeline (output/v2/) et les
                             JSONs legacy (output/).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Dom
2026-04-24 15:06:30 +02:00
parent 1f75670770
commit b6dd9ff1df
21 changed files with 9524 additions and 0 deletions

554
output/v2/OGC 74.json Normal file
View File

@@ -0,0 +1,554 @@
{
"fichier": "OGC 74",
"pdf_hash": "58076293464e9771",
"pages": [
{
"page": 1,
"type": "recueil",
"header": "Here's the transcription of the image:\n\nFICHE MEDICALE DE RECUEIL DU PRATICIEN CONSEIL (une fiche par RUM)\n(à transmettre au médecin du DIM avant concertation)\n\nSeul le recodage impactant la facturation est renseigné\n\nEtablissement : CLINIQUE D'ARCACHON\nFINESS : 330780206\nDate début contrôle : 13/02/2018\n\nN° champ : 1\nLibellé champ de contrôle : Champ 1 : séjours ayant des caractéristiques communes\nséjours de niveau 2, 3 et 4 mono RUM\nN° OGC : 74",
"elapsed_s": 8.02,
"ocr_raw": "```json\n{\n \"etablissement\": \"CLINIQUE D'ARCACHON\",\n \"finess\": \"330780206\",\n \"date_debut_controle\": \"13/02/2018\",\n \"n_ogc\": \"74\",\n \"n_champ\": \"1\",\n \"dates_sejour\": \"08/04/2016 au 13/04/2016\",\n \"sejour_etab\": {\n \"age\": \"52\", \"sexe\": \"1\", \"duree_sejour\": \"5\",\n \"mode_entree\": \"1\", \"provenance\": \"8\", \"mode_sortie\": \"1\", \"destination\": \"8\"\n },\n \"sejour_reco\": {\n \"age\": \"52\", \"sexe\": \"1\", \"duree_sejour\": \"5\",\n \"mode_entree\": \"1\", \"provenance\": \"8\", \"mode_sortie\": \"1\", \"destination\": \"8\"\n },\n \"rum_etab\": {\"um\": \"40 C\", \"igs\": \"II\", \"duree\": \"5\", \"dates\": \"du 08/04/2016 au 13/04/2016\"},\n \"codage_etab\": {\n \"dp\": \"A099\", \"dp_libelle\": \"GASTROENTERITE COLITE ORIGINE SAI\", \"dr\": \"C795 * 2\", \"das\": []\n },\n \"codage_reco\": {\n \"dp\": \"A099\", \"dr\": \"C795 * 2\", \"das\": []\n },\n \"actes_etab\": [], \"actes_reco\": [],\n \"ghm_etab\": \"06M032\", \"ghs_etab\": \"2130\",\n \"ghm_reco\": \"18M041\", \"ghs_reco\": \"6772\",\n \"recodage_impactant\": \"1\", \"ghs_injustifie\": \"0\", \"praticien_conseil\": \"DR JP VIGNAU\"\n}\n```",
"parsed": {
"etablissement": "CLINIQUE D'ARCACHON",
"finess": "330780206",
"date_debut_controle": "13/02/2018",
"n_ogc": "74",
"n_champ": "1",
"dates_sejour": "08/04/2016 au 13/04/2016",
"sejour_etab": {
"age": "52",
"sexe": "1",
"duree_sejour": "5",
"mode_entree": "1",
"provenance": "8",
"mode_sortie": "1",
"destination": "8"
},
"sejour_reco": {
"age": "52",
"sexe": "1",
"duree_sejour": "5",
"mode_entree": "1",
"provenance": "8",
"mode_sortie": "1",
"destination": "8"
},
"rum_etab": {
"um": "40 C",
"igs": "II",
"duree": "5",
"dates": "du 08/04/2016 au 13/04/2016"
},
"codage_etab": {
"dp": "A099",
"dp_libelle": "GASTROENTERITE COLITE ORIGINE SAI",
"dr": "C795 * 2",
"das": []
},
"codage_reco": {
"dp": "A099",
"dr": "C795 * 2",
"das": []
},
"actes_etab": [],
"actes_reco": [],
"ghm_etab": "06M032",
"ghs_etab": "2130",
"ghm_reco": "18M041",
"ghs_reco": "6772",
"recodage_impactant": "1",
"ghs_injustifie": "0",
"praticien_conseil": "DR JP VIGNAU",
"accord_desaccord": "désaccord",
"_checkbox_debug": {
"decision": "désaccord",
"ratio_accord": 0.1111,
"ratio_desaccord": 0.1422,
"diff": -0.0311
}
}
},
{
"page": 2,
"type": "concertation_med",
"header": "",
"elapsed_s": 0.0
},
{
"page": 3,
"type": "hospitalisation",
"header": "",
"elapsed_s": 0.0
},
{
"page": 4,
"type": "preuves",
"header": "",
"elapsed_s": 14.45,
"ocr_raw": "```json\n{\n \"date\": \"2023/01/18\",\n \"praticien_controleur\": [\"Dr RADZIKOWSKI\", \"Dr DELAYE-PHULPIN\", \"Dr TURBAN\", \"Dr DUVAL\", \"Dr VIGNAU\", \"Dr PROMAX\"],\n \"medecin_dim\": \"Dr ETTORCHI-TARDY\",\n \"pieces\": [\n {\n \"intitule\": \"Compte-rendu d'acte : TDP TAP TRN\",\n \"present\": true,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Compte-rendu opératoire\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Compte-rendu d'accouchement\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Compte-rendu d'examen complémentaire :\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Compte-rendu d'imagerie :\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Compte-rendu d'anatomopathologie\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Observations médicales\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Dossier de transfusion\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Dossier d'anesthésie\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Administration thérapeutique\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Compte-rendu d'hospitalisation\",\n \"present\": true,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Lettre de sortie\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Eléments de surveillance du dossier infirmier\",\n \"present\": true,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Prise en charge psychologique\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Prise en charge kinésithérapeute\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Prise en charge diététique\",\n \"present\": false,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n },\n {\n \"intitule\": \"Autre : b-c\",\n \"present\": true,\n \"photocopie\": false,\n \"absent_date\": \"\",\n \"date_obtention\": \"\"\n }\n ]\n}\n```",
"parsed": {
"date": "2023/01/18",
"praticien_controleur": [
"Dr RADZIKOWSKI",
"Dr DELAYE-PHULPIN",
"Dr TURBAN",
"Dr DUVAL",
"Dr VIGNAU",
"Dr PROMAX"
],
"medecin_dim": "Dr ETTORCHI-TARDY",
"pieces": [
{
"intitule": "Compte-rendu d'acte : TDP TAP TRN",
"present": true,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu opératoire",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'accouchement",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'examen complémentaire :",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'imagerie :",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'anatomopathologie",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Observations médicales",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Dossier de transfusion",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Dossier d'anesthésie",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Administration thérapeutique",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'hospitalisation",
"present": true,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Lettre de sortie",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Eléments de surveillance du dossier infirmier",
"present": true,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Prise en charge psychologique",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Prise en charge kinésithérapeute",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Prise en charge diététique",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Autre : b-c",
"present": true,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
}
]
}
},
{
"page": 5,
"type": "concertation_2",
"header": "",
"elapsed_s": 1.51,
"ocr_raw": "```json\n{\n \"ghs_initial\": \"\",\n \"ghs_avant_concertation\": \"\",\n \"ghs_final\": \"\",\n \"decision\": \"\",\n \"date_concertation\": \"2-3\",\n \"praticien_controleur\": \"\",\n \"medecin_dim\": \"\"\n}\n```",
"parsed": {
"ghs_initial": "",
"ghs_avant_concertation": "",
"ghs_final": "",
"decision": "",
"date_concertation": "2-3",
"praticien_controleur": "",
"medecin_dim": ""
}
},
{
"page": 6,
"type": "concertation_1",
"header": "",
"elapsed_s": 1.68,
"ocr_raw": "```json\n{\n \"date_concertation\": \"2.3.18\",\n \"argumentaire\": \"Atteste avoir pris connaissance des éléments du dossier y compris ceux couverts par le secret médical et des arguments soutenus par les médecins contrôleurs et avoir eu l'opportunité d'en débattre contradictoirement\"\n}\n```",
"parsed": {
"date_concertation": "2.3.18",
"argumentaire": "Atteste avoir pris connaissance des éléments du dossier y compris ceux couverts par le secret médical et des arguments soutenus par les médecins contrôleurs et avoir eu l'opportunité d'en débattre contradictoirement"
}
}
],
"extraction": {
"recueil": {
"etablissement": "CLINIQUE D'ARCACHON",
"finess": "330780206",
"date_debut_controle": "13/02/2018",
"n_ogc": "74",
"n_champ": "1",
"dates_sejour": "08/04/2016 au 13/04/2016",
"sejour_etab": {
"age": "52",
"sexe": "1",
"duree_sejour": "5",
"mode_entree": "1",
"provenance": "8",
"mode_sortie": "1",
"destination": "8"
},
"sejour_reco": {
"age": "52",
"sexe": "1",
"duree_sejour": "5",
"mode_entree": "1",
"provenance": "8",
"mode_sortie": "1",
"destination": "8"
},
"rum_etab": {
"um": "40 C",
"igs": "II",
"duree": "5",
"dates": "du 08/04/2016 au 13/04/2016"
},
"codage_etab": {
"dp": "A099",
"dp_libelle": "GASTROENTERITE COLITE ORIGINE SAI",
"dr": "C795 * 2",
"das": []
},
"codage_reco": {
"dp": "A099",
"dr": "C795 * 2",
"das": []
},
"actes_etab": [],
"actes_reco": [],
"ghm_etab": "06M032",
"ghs_etab": "2130",
"ghm_reco": "18M041",
"ghs_reco": "6772",
"recodage_impactant": "1",
"ghs_injustifie": "0",
"praticien_conseil": "DR JP VIGNAU",
"accord_desaccord": "désaccord",
"_checkbox_debug": {
"decision": "désaccord",
"ratio_accord": 0.1111,
"ratio_desaccord": 0.1422,
"diff": -0.0311
},
"_validation": {
"codage_etab": {
"dp": {
"code": "A099",
"valid": true,
"libelle_ref": "Gastroentérite et colite dorigine non précisée"
},
"dr": {
"code": "C795 * 2",
"valid": true,
"libelle_ref": "Tumeur maligne secondaire des os et de la moelle osseuse"
},
"das": []
},
"codage_reco": {
"dp": {
"code": "A099",
"valid": true,
"libelle_ref": "Gastroentérite et colite dorigine non précisée"
},
"dr": {
"code": "C795 * 2",
"valid": true,
"libelle_ref": "Tumeur maligne secondaire des os et de la moelle osseuse"
},
"das": []
},
"actes_etab": [],
"actes_reco": [],
"ghm_etab": {
"code": "06M032",
"valid": true,
"ghs_possibles": [
"2130"
]
},
"ghs_etab": {
"code": "2130",
"valid": true
},
"ghm_reco": {
"code": "18M041",
"valid": true,
"ghs_possibles": [
"6772"
]
},
"ghs_reco": {
"code": "6772",
"valid": true
},
"cross_checks": {
"etab": {
"checked": true,
"coherent": true,
"ghs_extrait": "2130",
"ghs_possibles": [
"2130"
]
},
"reco": {
"checked": true,
"coherent": true,
"ghs_extrait": "6772",
"ghs_possibles": [
"6772"
]
}
},
"summary": {
"valid": 8,
"invalid": 0,
"empty": 0,
"total_codes": 8,
"ghm_ghs_incoherents": 0
}
}
},
"preuves": {
"date": "2023/01/18",
"praticien_controleur": [
"Dr RADZIKOWSKI",
"Dr DELAYE-PHULPIN",
"Dr TURBAN",
"Dr DUVAL",
"Dr VIGNAU",
"Dr PROMAX"
],
"medecin_dim": "Dr ETTORCHI-TARDY",
"pieces": [
{
"intitule": "Compte-rendu d'acte : TDP TAP TRN",
"present": true,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu opératoire",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'accouchement",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'examen complémentaire :",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'imagerie :",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'anatomopathologie",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Observations médicales",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Dossier de transfusion",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Dossier d'anesthésie",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Administration thérapeutique",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Compte-rendu d'hospitalisation",
"present": true,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Lettre de sortie",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Eléments de surveillance du dossier infirmier",
"present": true,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Prise en charge psychologique",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Prise en charge kinésithérapeute",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Prise en charge diététique",
"present": false,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
},
{
"intitule": "Autre : b-c",
"present": true,
"photocopie": false,
"absent_date": "",
"date_obtention": ""
}
]
},
"concertation_2": {
"ghs_initial": "",
"ghs_avant_concertation": "",
"ghs_final": "",
"decision": "",
"date_concertation": "2-3",
"praticien_controleur": "",
"medecin_dim": "",
"_validation": {
"ghs_initial": {
"code": "",
"valid": null
},
"ghs_avant_concertation": {
"code": "",
"valid": null
},
"ghs_final": {
"code": "",
"valid": null
}
}
},
"concertation_1": {
"date_concertation": "2.3.18",
"argumentaire": "Atteste avoir pris connaissance des éléments du dossier y compris ceux couverts par le secret médical et des arguments soutenus par les médecins contrôleurs et avoir eu l'opportunité d'en débattre contradictoirement"
}
},
"_meta": {
"pipeline_version": "v1",
"ocr_model": "zai-org/GLM-OCR",
"generated_at": "2026-04-24T12:16:17+00:00"
}
}