mirror of
https://github.com/DS4SD/docling.git
synced 2025-12-08 20:58:11 +00:00
test: mark flaky test (#1698)
* test: cleanse Word test file Signed-off-by: Panos Vagenas <pva@zurich.ibm.com> * mark textbox file test as flaky Signed-off-by: Panos Vagenas <pva@zurich.ibm.com> * fix path usage Signed-off-by: Panos Vagenas <pva@zurich.ibm.com> --------- Signed-off-by: Panos Vagenas <pva@zurich.ibm.com>
This commit is contained in:
Panos Vagenas
GitHub
200
tests/data/groundtruth/docling_v2/textbox.docx.json
vendored
200
tests/data/groundtruth/docling_v2/textbox.docx.json
vendored
@@ -4,7 +4,7 @@
|
||||
"name": "textbox",
|
||||
"origin": {
|
||||
"mimetype": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
|
||||
"binary_hash": 830302052279341882,
|
||||
"binary_hash": 11723995438039370060,
|
||||
"filename": "textbox.docx"
|
||||
},
|
||||
"furniture": {
|
||||
@@ -66,7 +66,7 @@
|
||||
"$ref": "#/groups/4"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/22"
|
||||
"$ref": "#/groups/6"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/23"
|
||||
@@ -84,16 +84,16 @@
|
||||
"$ref": "#/texts/27"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/5"
|
||||
"$ref": "#/groups/7"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/7"
|
||||
"$ref": "#/groups/9"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/35"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/8"
|
||||
"$ref": "#/groups/10"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/37"
|
||||
@@ -117,7 +117,7 @@
|
||||
"$ref": "#/texts/43"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/9"
|
||||
"$ref": "#/groups/11"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/49"
|
||||
@@ -129,13 +129,13 @@
|
||||
"$ref": "#/texts/51"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/11"
|
||||
"$ref": "#/groups/13"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/55"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/12"
|
||||
"$ref": "#/groups/14"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/58"
|
||||
@@ -144,13 +144,13 @@
|
||||
"$ref": "#/texts/59"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/13"
|
||||
"$ref": "#/groups/15"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/60"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/14"
|
||||
"$ref": "#/groups/16"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/61"
|
||||
@@ -159,13 +159,13 @@
|
||||
"$ref": "#/texts/62"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/15"
|
||||
"$ref": "#/groups/17"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/67"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/16"
|
||||
"$ref": "#/groups/18"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/68"
|
||||
@@ -254,10 +254,7 @@
|
||||
},
|
||||
"children": [
|
||||
{
|
||||
"$ref": "#/texts/18"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/19"
|
||||
"$ref": "#/groups/5"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/20"
|
||||
@@ -272,6 +269,37 @@
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/5",
|
||||
"parent": {
|
||||
"$ref": "#/groups/4"
|
||||
},
|
||||
"children": [
|
||||
{
|
||||
"$ref": "#/texts/18"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/19"
|
||||
}
|
||||
],
|
||||
"content_layer": "body",
|
||||
"name": "list",
|
||||
"label": "list"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/6",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
"children": [
|
||||
{
|
||||
"$ref": "#/texts/22"
|
||||
}
|
||||
],
|
||||
"content_layer": "body",
|
||||
"name": "list",
|
||||
"label": "list"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/7",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
@@ -283,7 +311,7 @@
|
||||
"$ref": "#/texts/29"
|
||||
},
|
||||
{
|
||||
"$ref": "#/groups/6"
|
||||
"$ref": "#/groups/8"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/32"
|
||||
@@ -297,9 +325,9 @@
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/6",
|
||||
"self_ref": "#/groups/8",
|
||||
"parent": {
|
||||
"$ref": "#/groups/5"
|
||||
"$ref": "#/groups/7"
|
||||
},
|
||||
"children": [
|
||||
{
|
||||
@@ -314,7 +342,7 @@
|
||||
"label": "list"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/7",
|
||||
"self_ref": "#/groups/9",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
@@ -328,7 +356,7 @@
|
||||
"label": "list"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/8",
|
||||
"self_ref": "#/groups/10",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
@@ -342,13 +370,13 @@
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/9",
|
||||
"self_ref": "#/groups/11",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
"children": [
|
||||
{
|
||||
"$ref": "#/groups/10"
|
||||
"$ref": "#/groups/12"
|
||||
},
|
||||
{
|
||||
"$ref": "#/texts/47"
|
||||
@@ -362,9 +390,9 @@
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/10",
|
||||
"self_ref": "#/groups/12",
|
||||
"parent": {
|
||||
"$ref": "#/groups/9"
|
||||
"$ref": "#/groups/11"
|
||||
},
|
||||
"children": [
|
||||
{
|
||||
@@ -382,7 +410,7 @@
|
||||
"label": "inline"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/11",
|
||||
"self_ref": "#/groups/13",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
@@ -402,7 +430,7 @@
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/12",
|
||||
"self_ref": "#/groups/14",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
@@ -418,31 +446,31 @@
|
||||
"name": "textbox",
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/13",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
"name": "textbox",
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/14",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
"name": "textbox",
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/15",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
"name": "textbox",
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/16",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
"name": "textbox",
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/17",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
"children": [
|
||||
{
|
||||
"$ref": "#/texts/63"
|
||||
@@ -462,7 +490,7 @@
|
||||
"label": "section"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/groups/16",
|
||||
"self_ref": "#/groups/18",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
},
|
||||
@@ -732,38 +760,42 @@
|
||||
{
|
||||
"self_ref": "#/texts/18",
|
||||
"parent": {
|
||||
"$ref": "#/groups/4"
|
||||
"$ref": "#/groups/5"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
"label": "paragraph",
|
||||
"label": "list_item",
|
||||
"prov": [],
|
||||
"orig": " A report must be submitted within 24 hours via the Ministry of Education’s Campus Safety and Disaster Prevention Information Network.",
|
||||
"text": " A report must be submitted within 24 hours via the Ministry of Education’s Campus Safety and Disaster Prevention Information Network.",
|
||||
"orig": "A report must be submitted within 24 hours via the Ministry of Education’s Campus Safety and Disaster Prevention Information Network.",
|
||||
"text": "A report must be submitted within 24 hours via the Ministry of Education’s Campus Safety and Disaster Prevention Information Network.",
|
||||
"formatting": {
|
||||
"bold": false,
|
||||
"italic": false,
|
||||
"underline": false,
|
||||
"strikethrough": false
|
||||
}
|
||||
},
|
||||
"enumerated": false,
|
||||
"marker": "-"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/texts/19",
|
||||
"parent": {
|
||||
"$ref": "#/groups/4"
|
||||
"$ref": "#/groups/5"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
"label": "paragraph",
|
||||
"label": "list_item",
|
||||
"prov": [],
|
||||
"orig": " A report must also be submitted within 48 hours through Chiayi County’s School Suspected Infectious Disease Reporting System.",
|
||||
"text": " A report must also be submitted within 48 hours through Chiayi County’s School Suspected Infectious Disease Reporting System.",
|
||||
"orig": "A report must also be submitted within 48 hours through Chiayi County’s School Suspected Infectious Disease Reporting System.",
|
||||
"text": "A report must also be submitted within 48 hours through Chiayi County’s School Suspected Infectious Disease Reporting System.",
|
||||
"formatting": {
|
||||
"bold": false,
|
||||
"italic": false,
|
||||
"underline": false,
|
||||
"strikethrough": false
|
||||
}
|
||||
},
|
||||
"enumerated": false,
|
||||
"marker": "-"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/texts/20",
|
||||
@@ -792,14 +824,16 @@
|
||||
{
|
||||
"self_ref": "#/texts/22",
|
||||
"parent": {
|
||||
"$ref": "#/body"
|
||||
"$ref": "#/groups/6"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
"label": "paragraph",
|
||||
"label": "list_item",
|
||||
"prov": [],
|
||||
"orig": "",
|
||||
"text": ""
|
||||
"text": "",
|
||||
"enumerated": false,
|
||||
"marker": "-"
|
||||
},
|
||||
{
|
||||
"self_ref": "#/texts/23",
|
||||
@@ -864,7 +898,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/28",
|
||||
"parent": {
|
||||
"$ref": "#/groups/5"
|
||||
"$ref": "#/groups/7"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -882,7 +916,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/29",
|
||||
"parent": {
|
||||
"$ref": "#/groups/5"
|
||||
"$ref": "#/groups/7"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -900,7 +934,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/30",
|
||||
"parent": {
|
||||
"$ref": "#/groups/6"
|
||||
"$ref": "#/groups/8"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -920,7 +954,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/31",
|
||||
"parent": {
|
||||
"$ref": "#/groups/6"
|
||||
"$ref": "#/groups/8"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -940,7 +974,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/32",
|
||||
"parent": {
|
||||
"$ref": "#/groups/5"
|
||||
"$ref": "#/groups/7"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -952,7 +986,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/33",
|
||||
"parent": {
|
||||
"$ref": "#/groups/5"
|
||||
"$ref": "#/groups/7"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -964,7 +998,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/34",
|
||||
"parent": {
|
||||
"$ref": "#/groups/7"
|
||||
"$ref": "#/groups/9"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -990,7 +1024,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/36",
|
||||
"parent": {
|
||||
"$ref": "#/groups/8"
|
||||
"$ref": "#/groups/10"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1092,7 +1126,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/44",
|
||||
"parent": {
|
||||
"$ref": "#/groups/10"
|
||||
"$ref": "#/groups/12"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1110,7 +1144,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/45",
|
||||
"parent": {
|
||||
"$ref": "#/groups/10"
|
||||
"$ref": "#/groups/12"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1128,7 +1162,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/46",
|
||||
"parent": {
|
||||
"$ref": "#/groups/10"
|
||||
"$ref": "#/groups/12"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1146,7 +1180,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/47",
|
||||
"parent": {
|
||||
"$ref": "#/groups/9"
|
||||
"$ref": "#/groups/11"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1158,7 +1192,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/48",
|
||||
"parent": {
|
||||
"$ref": "#/groups/9"
|
||||
"$ref": "#/groups/11"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1206,7 +1240,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/52",
|
||||
"parent": {
|
||||
"$ref": "#/groups/11"
|
||||
"$ref": "#/groups/13"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1224,7 +1258,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/53",
|
||||
"parent": {
|
||||
"$ref": "#/groups/11"
|
||||
"$ref": "#/groups/13"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1236,7 +1270,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/54",
|
||||
"parent": {
|
||||
"$ref": "#/groups/11"
|
||||
"$ref": "#/groups/13"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1260,7 +1294,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/56",
|
||||
"parent": {
|
||||
"$ref": "#/groups/12"
|
||||
"$ref": "#/groups/14"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1278,7 +1312,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/57",
|
||||
"parent": {
|
||||
"$ref": "#/groups/12"
|
||||
"$ref": "#/groups/14"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1356,7 +1390,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/63",
|
||||
"parent": {
|
||||
"$ref": "#/groups/15"
|
||||
"$ref": "#/groups/17"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1374,7 +1408,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/64",
|
||||
"parent": {
|
||||
"$ref": "#/groups/15"
|
||||
"$ref": "#/groups/17"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1386,7 +1420,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/65",
|
||||
"parent": {
|
||||
"$ref": "#/groups/15"
|
||||
"$ref": "#/groups/17"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
@@ -1398,7 +1432,7 @@
|
||||
{
|
||||
"self_ref": "#/texts/66",
|
||||
"parent": {
|
||||
"$ref": "#/groups/15"
|
||||
"$ref": "#/groups/17"
|
||||
},
|
||||
"children": [],
|
||||
"content_layer": "body",
|
||||
|
||||
Reference in New Issue
Block a user