Adds option to do not abort the query when highlighting #3842
Labels
enhancement
Enhancement or improvement to existing feature or request
feedback needed
Issue or PR needs feedback
Indexing & Search
v2.2.0
v3.0.0
Issues and PRs related to version 3.0.0
Comments
hauck-jvsh
added
enhancement
mch2
added
feedback needed
|
Do you guys needs more info? I can link the Elasticsearch issue here or it is forbidden? |
|
You should feel free to link issues. If you want to contribute a fix, and if ES has fixed this issue, please feel free to reimplement and confirm that you have not been looking at any non-APLv2-compatible-licensed code. |
|
This issue was fixed in Elasticsearch 7.12, the issue and the pull request is available in the links bellow: |
hauck-jvsh
added a commit
to hauck-jvsh/OpenSearch
that referenced
this issue
Jul 13, 2022
…max_analyzer_offset. When this parameter is provided the highlight stops in its value. This prevents the highlighter to go beyond the index maxAnalyzedOffset.
5 tasks
hauck-jvsh
added a commit
to hauck-jvsh/OpenSearch
that referenced
this issue
Jul 15, 2022
…max_analyzer_offset. When this parameter is provided the highlight stops in its value. This prevents the highlighter to go beyond the index maxAnalyzedOffset. Signed-off-by: Hauck <joaoh14@gmail.com>
andrross
pushed a commit
that referenced
this issue
Jul 22, 2022
* #3842 adds a new parameter to the highlighter, the max_analyzer_offset. When this parameter is provided the highlight stops in its value. This prevents the highlighter to go beyond the index maxAnalyzedOffset. Signed-off-by: Hauck <joaoh14@gmail.com> * Adds a test for the new parameter Signed-off-by: Hauck <joaoh14@gmail.com> * Fix the test add in the previous commit; Signed-off-by: Hauck <joaoh14@gmail.com> * This was checking against the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Only runs the test for the correct version Signed-off-by: Hauck <joaoh14@gmail.com> * Skips the test in Elasticsearch as well; Signed-off-by: Hauck <joaoh14@gmail.com> * Remove elastic 3.0 to test Signed-off-by: Hauck <joaoh14@gmail.com> * Skips all versions Signed-off-by: Hauck <joaoh14@gmail.com> * Remove unnecessary fields as pointed by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * Compute if fieldMaxAnalyzedIsNotValid in the constructor as suggest by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * As discussed, it is better to throws different exceptions for when the fieldMaxAnalyzed is not valid and for when it is disabled; Signed-off-by: Hauck <joaoh14@gmail.com> * hint what to do to allow highlight of bigger documents Signed-off-by: Hauck <joaoh14@gmail.com> * Let the user define the new parameter globally for all fields highlighted Signed-off-by: Hauck <joaoh14@gmail.com> * Change the fieldMaxAnalyzedOffset Integer in order to use null when it is absent in highlight. This allows the error messages to much more precise, showing invalid for all negative numbers; Signed-off-by: Hauck <joaoh14@gmail.com> * Update javadocs and implements the stream methods for the new fields; Signed-off-by: Hauck <joaoh14@gmail.com> * builder.field do not accept null, so check before calling the method is necessary Signed-off-by: Hauck <joaoh14@gmail.com> * Only send and read the new fields if the version supports it Signed-off-by: Hauck <joaoh14@gmail.com> * the previous commit was checking the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Check for version 3.0.0 instead of current version Signed-off-by: Hauck <joaoh14@gmail.com> * Update server/src/main/java/org/apache/lucene/search/uhighlight/CustomUnifiedHighlighter.java Co-authored-by: Andriy Redko <drreta@gmail.com> Signed-off-by: Hauck <joaoh14@gmail.com> * Execute the test after version 3.0.0 Signed-off-by: Hauck <joaoh14@gmail.com> Co-authored-by: Andriy Redko <drreta@gmail.com>
opensearch-trigger-bot bot
pushed a commit
that referenced
this issue
Jul 22, 2022
* #3842 adds a new parameter to the highlighter, the max_analyzer_offset. When this parameter is provided the highlight stops in its value. This prevents the highlighter to go beyond the index maxAnalyzedOffset. Signed-off-by: Hauck <joaoh14@gmail.com> * Adds a test for the new parameter Signed-off-by: Hauck <joaoh14@gmail.com> * Fix the test add in the previous commit; Signed-off-by: Hauck <joaoh14@gmail.com> * This was checking against the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Only runs the test for the correct version Signed-off-by: Hauck <joaoh14@gmail.com> * Skips the test in Elasticsearch as well; Signed-off-by: Hauck <joaoh14@gmail.com> * Remove elastic 3.0 to test Signed-off-by: Hauck <joaoh14@gmail.com> * Skips all versions Signed-off-by: Hauck <joaoh14@gmail.com> * Remove unnecessary fields as pointed by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * Compute if fieldMaxAnalyzedIsNotValid in the constructor as suggest by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * As discussed, it is better to throws different exceptions for when the fieldMaxAnalyzed is not valid and for when it is disabled; Signed-off-by: Hauck <joaoh14@gmail.com> * hint what to do to allow highlight of bigger documents Signed-off-by: Hauck <joaoh14@gmail.com> * Let the user define the new parameter globally for all fields highlighted Signed-off-by: Hauck <joaoh14@gmail.com> * Change the fieldMaxAnalyzedOffset Integer in order to use null when it is absent in highlight. This allows the error messages to much more precise, showing invalid for all negative numbers; Signed-off-by: Hauck <joaoh14@gmail.com> * Update javadocs and implements the stream methods for the new fields; Signed-off-by: Hauck <joaoh14@gmail.com> * builder.field do not accept null, so check before calling the method is necessary Signed-off-by: Hauck <joaoh14@gmail.com> * Only send and read the new fields if the version supports it Signed-off-by: Hauck <joaoh14@gmail.com> * the previous commit was checking the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Check for version 3.0.0 instead of current version Signed-off-by: Hauck <joaoh14@gmail.com> * Update server/src/main/java/org/apache/lucene/search/uhighlight/CustomUnifiedHighlighter.java Co-authored-by: Andriy Redko <drreta@gmail.com> Signed-off-by: Hauck <joaoh14@gmail.com> * Execute the test after version 3.0.0 Signed-off-by: Hauck <joaoh14@gmail.com> Co-authored-by: Andriy Redko <drreta@gmail.com> (cherry picked from commit 931813f)
hauck-jvsh
added a commit
to hauck-jvsh/OpenSearch
that referenced
this issue
Jul 28, 2022
…earch-project#3893) * opensearch-project#3842 adds a new parameter to the highlighter, the max_analyzer_offset. When this parameter is provided the highlight stops in its value. This prevents the highlighter to go beyond the index maxAnalyzedOffset. Signed-off-by: Hauck <joaoh14@gmail.com> * Adds a test for the new parameter Signed-off-by: Hauck <joaoh14@gmail.com> * Fix the test add in the previous commit; Signed-off-by: Hauck <joaoh14@gmail.com> * This was checking against the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Only runs the test for the correct version Signed-off-by: Hauck <joaoh14@gmail.com> * Skips the test in Elasticsearch as well; Signed-off-by: Hauck <joaoh14@gmail.com> * Remove elastic 3.0 to test Signed-off-by: Hauck <joaoh14@gmail.com> * Skips all versions Signed-off-by: Hauck <joaoh14@gmail.com> * Remove unnecessary fields as pointed by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * Compute if fieldMaxAnalyzedIsNotValid in the constructor as suggest by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * As discussed, it is better to throws different exceptions for when the fieldMaxAnalyzed is not valid and for when it is disabled; Signed-off-by: Hauck <joaoh14@gmail.com> * hint what to do to allow highlight of bigger documents Signed-off-by: Hauck <joaoh14@gmail.com> * Let the user define the new parameter globally for all fields highlighted Signed-off-by: Hauck <joaoh14@gmail.com> * Change the fieldMaxAnalyzedOffset Integer in order to use null when it is absent in highlight. This allows the error messages to much more precise, showing invalid for all negative numbers; Signed-off-by: Hauck <joaoh14@gmail.com> * Update javadocs and implements the stream methods for the new fields; Signed-off-by: Hauck <joaoh14@gmail.com> * builder.field do not accept null, so check before calling the method is necessary Signed-off-by: Hauck <joaoh14@gmail.com> * Only send and read the new fields if the version supports it Signed-off-by: Hauck <joaoh14@gmail.com> * the previous commit was checking the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Check for version 3.0.0 instead of current version Signed-off-by: Hauck <joaoh14@gmail.com> * Update server/src/main/java/org/apache/lucene/search/uhighlight/CustomUnifiedHighlighter.java Co-authored-by: Andriy Redko <drreta@gmail.com> Signed-off-by: Hauck <joaoh14@gmail.com> * Execute the test after version 3.0.0 Signed-off-by: Hauck <joaoh14@gmail.com> Co-authored-by: Andriy Redko <drreta@gmail.com> (cherry picked from commit 931813f) Signed-off-by: Hauck <joaoh14@gmail.com>
reta
added a commit
that referenced
this issue
Jul 28, 2022
…hlighter (#4031) * Adds a new parameter, max_analyzer_offset, for the highlighter (#3893) * #3842 adds a new parameter to the highlighter, the max_analyzer_offset. When this parameter is provided the highlight stops in its value. This prevents the highlighter to go beyond the index maxAnalyzedOffset. Signed-off-by: Hauck <joaoh14@gmail.com> * Adds a test for the new parameter Signed-off-by: Hauck <joaoh14@gmail.com> * Fix the test add in the previous commit; Signed-off-by: Hauck <joaoh14@gmail.com> * This was checking against the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Only runs the test for the correct version Signed-off-by: Hauck <joaoh14@gmail.com> * Skips the test in Elasticsearch as well; Signed-off-by: Hauck <joaoh14@gmail.com> * Remove elastic 3.0 to test Signed-off-by: Hauck <joaoh14@gmail.com> * Skips all versions Signed-off-by: Hauck <joaoh14@gmail.com> * Remove unnecessary fields as pointed by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * Compute if fieldMaxAnalyzedIsNotValid in the constructor as suggest by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * As discussed, it is better to throws different exceptions for when the fieldMaxAnalyzed is not valid and for when it is disabled; Signed-off-by: Hauck <joaoh14@gmail.com> * hint what to do to allow highlight of bigger documents Signed-off-by: Hauck <joaoh14@gmail.com> * Let the user define the new parameter globally for all fields highlighted Signed-off-by: Hauck <joaoh14@gmail.com> * Change the fieldMaxAnalyzedOffset Integer in order to use null when it is absent in highlight. This allows the error messages to much more precise, showing invalid for all negative numbers; Signed-off-by: Hauck <joaoh14@gmail.com> * Update javadocs and implements the stream methods for the new fields; Signed-off-by: Hauck <joaoh14@gmail.com> * builder.field do not accept null, so check before calling the method is necessary Signed-off-by: Hauck <joaoh14@gmail.com> * Only send and read the new fields if the version supports it Signed-off-by: Hauck <joaoh14@gmail.com> * the previous commit was checking the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Check for version 3.0.0 instead of current version Signed-off-by: Hauck <joaoh14@gmail.com> * Update server/src/main/java/org/apache/lucene/search/uhighlight/CustomUnifiedHighlighter.java Co-authored-by: Andriy Redko <drreta@gmail.com> Signed-off-by: Hauck <joaoh14@gmail.com> * Execute the test after version 3.0.0 Signed-off-by: Hauck <joaoh14@gmail.com> Co-authored-by: Andriy Redko <drreta@gmail.com> (cherry picked from commit 931813f) Signed-off-by: Hauck <joaoh14@gmail.com> * Fix the version tests Signed-off-by: Hauck <joaoh14@gmail.com> Co-authored-by: Andriy Redko <drreta@gmail.com>
dblock
pushed a commit
that referenced
this issue
Nov 8, 2022
* #3842 adds a new parameter to the highlighter, the max_analyzer_offset. When this parameter is provided the highlight stops in its value. This prevents the highlighter to go beyond the index maxAnalyzedOffset. Signed-off-by: Hauck <joaoh14@gmail.com> * Adds a test for the new parameter Signed-off-by: Hauck <joaoh14@gmail.com> * Fix the test add in the previous commit; Signed-off-by: Hauck <joaoh14@gmail.com> * This was checking against the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Only runs the test for the correct version Signed-off-by: Hauck <joaoh14@gmail.com> * Skips the test in Elasticsearch as well; Signed-off-by: Hauck <joaoh14@gmail.com> * Remove elastic 3.0 to test Signed-off-by: Hauck <joaoh14@gmail.com> * Skips all versions Signed-off-by: Hauck <joaoh14@gmail.com> * Remove unnecessary fields as pointed by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * Compute if fieldMaxAnalyzedIsNotValid in the constructor as suggest by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * As discussed, it is better to throws different exceptions for when the fieldMaxAnalyzed is not valid and for when it is disabled; Signed-off-by: Hauck <joaoh14@gmail.com> * hint what to do to allow highlight of bigger documents Signed-off-by: Hauck <joaoh14@gmail.com> * Let the user define the new parameter globally for all fields highlighted Signed-off-by: Hauck <joaoh14@gmail.com> * Change the fieldMaxAnalyzedOffset Integer in order to use null when it is absent in highlight. This allows the error messages to much more precise, showing invalid for all negative numbers; Signed-off-by: Hauck <joaoh14@gmail.com> * Update javadocs and implements the stream methods for the new fields; Signed-off-by: Hauck <joaoh14@gmail.com> * builder.field do not accept null, so check before calling the method is necessary Signed-off-by: Hauck <joaoh14@gmail.com> * Only send and read the new fields if the version supports it Signed-off-by: Hauck <joaoh14@gmail.com> * the previous commit was checking the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Check for version 3.0.0 instead of current version Signed-off-by: Hauck <joaoh14@gmail.com> * Update server/src/main/java/org/apache/lucene/search/uhighlight/CustomUnifiedHighlighter.java Co-authored-by: Andriy Redko <drreta@gmail.com> Signed-off-by: Hauck <joaoh14@gmail.com> * Execute the test after version 3.0.0 Signed-off-by: Hauck <joaoh14@gmail.com> Co-authored-by: Andriy Redko <drreta@gmail.com> (cherry picked from commit 931813f)
dblock
pushed a commit
that referenced
this issue
Nov 8, 2022
* #3842 adds a new parameter to the highlighter, the max_analyzer_offset. When this parameter is provided the highlight stops in its value. This prevents the highlighter to go beyond the index maxAnalyzedOffset. Signed-off-by: Hauck <joaoh14@gmail.com> * Adds a test for the new parameter Signed-off-by: Hauck <joaoh14@gmail.com> * Fix the test add in the previous commit; Signed-off-by: Hauck <joaoh14@gmail.com> * This was checking against the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Only runs the test for the correct version Signed-off-by: Hauck <joaoh14@gmail.com> * Skips the test in Elasticsearch as well; Signed-off-by: Hauck <joaoh14@gmail.com> * Remove elastic 3.0 to test Signed-off-by: Hauck <joaoh14@gmail.com> * Skips all versions Signed-off-by: Hauck <joaoh14@gmail.com> * Remove unnecessary fields as pointed by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * Compute if fieldMaxAnalyzedIsNotValid in the constructor as suggest by @reta Signed-off-by: Hauck <joaoh14@gmail.com> * As discussed, it is better to throws different exceptions for when the fieldMaxAnalyzed is not valid and for when it is disabled; Signed-off-by: Hauck <joaoh14@gmail.com> * hint what to do to allow highlight of bigger documents Signed-off-by: Hauck <joaoh14@gmail.com> * Let the user define the new parameter globally for all fields highlighted Signed-off-by: Hauck <joaoh14@gmail.com> * Change the fieldMaxAnalyzedOffset Integer in order to use null when it is absent in highlight. This allows the error messages to much more precise, showing invalid for all negative numbers; Signed-off-by: Hauck <joaoh14@gmail.com> * Update javadocs and implements the stream methods for the new fields; Signed-off-by: Hauck <joaoh14@gmail.com> * builder.field do not accept null, so check before calling the method is necessary Signed-off-by: Hauck <joaoh14@gmail.com> * Only send and read the new fields if the version supports it Signed-off-by: Hauck <joaoh14@gmail.com> * the previous commit was checking the wrong field Signed-off-by: Hauck <joaoh14@gmail.com> * Check for version 3.0.0 instead of current version Signed-off-by: Hauck <joaoh14@gmail.com> * Update server/src/main/java/org/apache/lucene/search/uhighlight/CustomUnifiedHighlighter.java Co-authored-by: Andriy Redko <drreta@gmail.com> Signed-off-by: Hauck <joaoh14@gmail.com> * Execute the test after version 3.0.0 Signed-off-by: Hauck <joaoh14@gmail.com> Co-authored-by: Andriy Redko <drreta@gmail.com> (cherry picked from commit 931813f)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I have an index with very large texts inside a field, and when a search with highlighting is made against it, the following error is returned:
I think that it will be better if the query can execute and only the highlight is disabled. Another option is to add an max_analyzed_offset option to the highlighter and ignore all after this option similar to what is done in the current Elasticsearch version.
The text was updated successfully, but these errors were encountered: