Merge pull request #1068 from kermitt2/feature/paragraphs-coordinates

Add paragraphs coordinates
kermitt2 · Dec 29, 2023 · b50c9aa · b50c9aa
2 parents 2ca3f35 + a506eb1
commit b50c9aa
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 3 deletions.
diff --git a/doc/Coordinates-in-PDF.md b/doc/Coordinates-in-PDF.md
@@ -13,6 +13,7 @@ Since April 2017, GROBID version 0.4.2 and higher, coordinate areas can be obtai
 * ```formula``` for mathematical equations,
 * ```head``` for section titles,
 * ```s``` for optional sentence structure (the GROBID fulltext service must be called with the `segmentSentences` parameter to provide the optional sentence-level elements),
+* ```p``` for paragraph structure,
 * ```note``` for foot note elements,
 * ```title``` for the title elements (main article title and cited reference titles),
 * ```affiliation``` for the affiliation and address part.
@@ -117,7 +118,7 @@ The GROBID console offers a reference implementation with PDF.js for dynamically
 
 ### Coordinates in TEI/XML results
 
-Coordinates for a given structure appear via an extra attribute ```@coord```. This is part of the [customization to the TEI](TEI-encoding-of-results.md) used by GROBID.
+Coordinates for a given structure appear via an extra attribute ```@coords```. This is part of the [customization to the TEI](TEI-encoding-of-results.md) used by GROBID.
 
 * the list of page size is encoded under the TEI element `<facsimile>`. The dimension of each page is given successively by the TEI attributes `@lrx` and `@lry` of the element `<surface>` to be conformant with the TEI (`@ulx` and `@uly` are used to set the orgine coordinates, which is always `(0,0)` for us).
 

diff --git a/grobid-core/src/main/java/org/grobid/core/document/TEIFormatter.java b/grobid-core/src/main/java/org/grobid/core/document/TEIFormatter.java
@@ -1202,7 +1202,12 @@ private StringBuilder toTEINote(StringBuilder tei,
                 String pID = KeyGen.getKey().substring(0, 7);
                 addXmlId(pNote, "_" + pID);
             }
-
+
+            if (config.isGenerateTeiCoordinates("p")) {
+                String coords = LayoutTokensUtil.getCoordsString(note.getTokens());
+                desc.addAttribute(new Attribute("coords", coords));
+            }
+
             // for labelling bibliographical references in notes 
             List<LayoutToken> noteTokens = note.getTokens();
 
@@ -1472,6 +1477,12 @@ public StringBuilder toTEITextPiece(StringBuilder buffer,
                             String divID = KeyGen.getKey().substring(0, 7);
                             addXmlId(curParagraph, "_" + divID);
                         }
+
+                        if (config.isGenerateTeiCoordinates("p")) {
+                            String coords = LayoutTokensUtil.getCoordsString(clusterTokens);
+                            curParagraph.addAttribute(new Attribute("coords", coords));
+                        }
+
                         curDiv.appendChild(curParagraph);
                         curParagraphTokens = new ArrayList<>();
                     }
@@ -1487,6 +1498,12 @@ public StringBuilder toTEITextPiece(StringBuilder buffer,
                             String divID = KeyGen.getKey().substring(0, 7);
                             addXmlId(curParagraph, "_" + divID);
                         }
+
+                        if (config.isGenerateTeiCoordinates("p")) {
+                            String coords = LayoutTokensUtil.getCoordsString(clusterTokens);
+                            curParagraph.addAttribute(new Attribute("coords", coords));
+                        }
+
                         curDiv.appendChild(curParagraph);
                         curParagraphTokens = new ArrayList<>();
                     }
@@ -1543,6 +1560,13 @@ public StringBuilder toTEITextPiece(StringBuilder buffer,
                         }
 
                         curParagraph.appendChild(clusterContentBefore);
+                        if (config.isGenerateTeiCoordinates("p")) {
+                            String coords = LayoutTokensUtil.getCoordsString(before);
+                            if (curParagraph.getAttribute("coords") != null && !curParagraph.getAttributeValue("coords").contains(coords)) {
+                                curParagraph.addAttribute(new Attribute("coords", curParagraph.getAttributeValue("coords") + ";" + coords));
+                            }
+                        }
+
                         curParagraphTokens.addAll(before);
 
                         List<LayoutToken> calloutTokens = clusterTokens.subList(matchingPosition.start, matchingPosition.end);
@@ -1572,6 +1596,13 @@ public StringBuilder toTEITextPiece(StringBuilder buffer,
                         curParagraph.appendChild(new Text(" "));
                     }
 
+                    if (config.isGenerateTeiCoordinates("p")) {
+                        String coords = LayoutTokensUtil.getCoordsString(remaining);
+                        if (curParagraph.getAttribute("coords") != null && !curParagraph.getAttributeValue("coords").contains(coords)) {
+                            curParagraph.addAttribute(new Attribute("coords", curParagraph.getAttributeValue("coords") + ";" + coords));
+                        }
+                    }
+
                     curParagraph.appendChild(remainingClusterContent);
                     curParagraphTokens.addAll(remaining);
                 }

diff --git a/grobid-core/src/main/java/org/grobid/core/engines/ProcessEngine.java b/grobid-core/src/main/java/org/grobid/core/engines/ProcessEngine.java
@@ -137,7 +137,7 @@ public void processFullText(final GrobidMainArgs pGbdArgs) throws Exception {
         } else {
             List<String> elementCoordinates = null;
             if (pGbdArgs.getTeiCoordinates()) {
-                elementCoordinates = Arrays.asList("figure", "persName", "ref", "biblStruct", "formula", "s", "note", "title", "head", "affiliation");
+                elementCoordinates = Arrays.asList("figure", "persName", "ref", "biblStruct", "formula", "s", "note", "title", "head", "affiliation", "p");
             }
             processFullTextDirectory(files, pGbdArgs, pGbdArgs.getPath2Output(), pGbdArgs.getSaveAssets(), 
                 elementCoordinates, pGbdArgs.getSegmentSentences(), pGbdArgs.getAddElementId());