Feature/enable paginated fetchers #7082
Conversation
Signed-off-by: Dominik Voigt <[email protected]>
Signed-off-by: Dominik Voigt <[email protected]>
| // if there are too much requests from the same IP address google is answering with a 503 and redirecting to a captcha challenge | ||
| // The caught IOException looks for example like this: | ||
| // java.io.IOException: Server returned HTTP response code: 503 for URL: https://ipv4.google.com/sorry/index?continue=https://scholar.google.com/scholar%3Fhl%3Den%26btnG%3DSearch%26q%3Dbpmn&hl=en&q=CGMSBI0NBDkYuqy9wAUiGQDxp4NLQCWbIEY1HjpH5zFJhv4ANPGdWj0 | ||
| if (e.getMessage().contains("Server returned HTTP response code: 503 for URL")) { |
There was a problem hiding this comment.
This struck me recently, isn't there a way to check if the exception is a more specific exception so that one could check for the status code number?
There was a problem hiding this comment.
I do not quite understand what the benefits of that would be.
Could you provide me with an example to grasp your idea, please? :)
Signed-off-by: Dominik Voigt <[email protected]>
There was a problem hiding this comment.
You have currently quite a lot of code duplications. The main sources are:
- Complex and string-based queries have to be implemented separately, leading to mostly the same code. Suggestion: remove string-based queries completely
- Unpaged search methods always fall back to paged search method with page number 0. Suggestion: put these fallbacks in the general
PagedSearchFetcherinterface.
| @@ -234,8 +228,31 @@ public String getName() { | |||
|
|
|||
| @Override | |||
| public URL getComplexQueryURL(ComplexSearchQuery complexSearchQuery) throws URISyntaxException, MalformedURLException { | |||
| return getComplexQueryURL(complexSearchQuery, 0); | |||
There was a problem hiding this comment.
This should be the default implementation, i.e. push to PagedSearchBasedParserFetcher
There was a problem hiding this comment.
You have currently quite a lot of code duplications. The main sources are:
- Complex and string-based queries have to be implemented separately, leading to mostly the same code. Suggestion: remove string-based queries completely
- Unpaged search methods always fall back to paged search method with page number 0. Suggestion: put these fallbacks in the general
PagedSearchFetcherinterface.
Thanks for your feedback! :)
Regarding 1:
I think that this is a good idea but will require more work than just replacing the method.
This is because for the WebSearchPane for instance the normal string-based search is used as a fallback case when the query could not be parsed.
Therefore I do not feel comfortable just tossing the string-based version.
However, I will address this in the upcoming weeks :)!
There was a problem hiding this comment.
I implemented all other suggestions
There was a problem hiding this comment.
Good!
Concerning the first point, doesn't it work to parse the queries as follows:
author=me and title=something -> ComplexQuery[author: "me", title: "something", rest: ""]
author=me something -> ComplexQuery[author: "me", title: "", rest: "something"]
something -> ComplexQuery[author: "", title: "", rest: ""something]
Then you don't need any fall-back to a purely string-based search.
In my opinion, this question should be resolved before changing the fetcher in other ways. Otherwise you have a lot of overhead/code duplication now, that will be removed later.
There was a problem hiding this comment.
I now removed the normal performSearch all together :)
There was a problem hiding this comment.
Thanks a lot! If you now also remove the "complex" in the names, I'm super happy ;-)
There was a problem hiding this comment.
I now removed the complex from perform search :)
Regarding the query object, I believe that it is sensible to keep that name as it will be extended in the future such as structured information.
src/test/java/org/jabref/logic/importer/fetcher/SpringerFetcherTest.java
Outdated
Show resolved
Hide resolved
Move common methods into default implementation Signed-off-by: Dominik Voigt <[email protected]>
Signed-off-by: Dominik Voigt <[email protected]>
Signed-off-by: Dominik Voigt <[email protected]>
Adapt query parser and add new tests Signed-off-by: Dominik Voigt <[email protected]>
This PR adds complex search query support for paginated fetchers.
It additionally implements the corresponding interfaces for a couple of fetchers (arXiv, Scholar, IEEE Xplore, Springer Link).
Refs #6236, #5507, koppor#369, koppor#347