[WIP] Rhar768/feature/23 rest interface #59
Conversation
Code will need refactoring in terms of its package once parser and REST interface set up.
…text-mindmap-parser
…/38-bibtext-mindmap-parser
This commit also fixes code style issues
rowanh22
changed the title
There was a problem hiding this comment.
Minor comments.
Please ensure that CheckStyle (and IntelliJ) is correctly setup at your side. Then, you can press Ctrl+Alt+L to fix the formatting (and Ctrl+Alt+O "Optimize Imports). One "annoying" thing of JabRef's checkstyle config is that there is a space needed after comment start (OK: // comment, not OK: //comment)
If possible, add Tests for the serialization. One can press Ctrl+Shift+T to quickly create test case skelletons.
See org.jabref.logic.bibtex.BibEntryWriterTest for an example test for BibEntry writing.
| import java.util.ArrayList; | ||
| import java.util.List; | ||
|
|
||
| @Path("/root") |
There was a problem hiding this comment.
Is it possible to omit /root?
| @Path("/root") | |
| @Path("/") |
| public class RootResource { | ||
|
|
||
| @GET | ||
| @Path("entries") |
There was a problem hiding this comment.
Please foresee the possibilitiy to work on multiple libraries (libraries/current)
| @Path("entries") | |
| @Path("libraries/current/entries") |
| @Path("entries") | ||
| @Produces(MediaType.APPLICATION_JSON) | ||
| public String getEntries() { | ||
| List<BibEntry> entries = MindMapManager.instance().getBibEntries(); |
There was a problem hiding this comment.
This code is a bit strange. I would go to the current opened library directly instead of routing the request throught he MindMapManager. (Since one does not need to be that specific here, the more general "object" is enough)
| public JsonElement serialize(BibEntry entry, Type typeOfSrc, JsonSerializationContext context) { | ||
| JsonObject obj = new JsonObject(); | ||
|
|
||
| obj.addProperty("type", entry.getType().getName()); |
There was a problem hiding this comment.
Please introduce a constant (public static final String JSON_TYPE)
General thing: We need to add two meta information to the JSON. We need to the bibtex content and the meta data type and key. They should not be mixed up. We need to consider that a user can add custom fields.
To be prepared for moree meta data, I would suggest a nested object
{
"bibtex_metadata": {
"type": "article",
"key": "the-citation-key"
}
}In case a user adds bibtex_metadata as custom field in the entry editor, then the thing breaks.
Alternative: Do not add special handling for citationkey, because JabRef itself uses that field name.
Another thing:
Shouldn't one use typedEntry.getTypeForDisplay() for type writing (Source: org.jabref.logic.bibtex.BibEntryWriter#writeRequiredFieldsFirstRemainingFieldsSecond)
TypedBibEntry typedEntry = new TypedBibEntry(entry, bibDatabaseMode);
typedEntry.getTypeForDisplay();| JsonObject obj = new JsonObject(); | ||
|
|
||
| obj.addProperty("type", entry.getType().getName()); | ||
| obj.addProperty("cite_key", entry.getCiteKeyOptional().orElse("")); |
There was a problem hiding this comment.
Difficult decision here. Alternatives are null or to omit the field. It is OK for me to use the emptry string. Seems to be the JavaScript way.
MindMapNodes no longer keep list of children but have x and y pos. Done to prevent cyclic dependency complications when serialising.
…ature/23-REST-interface # Conflicts: # src/main/java/org/jabref/gui/jabmap/JabMapAction.java # src/main/java/org/jabref/logic/jabmap/MindMapManager.java # src/main/java/org/jabref/rest/resources/RootResource.java
|
Follow-up at #76. |
Refs #23