Build Diagnostics

Diagnostics

The values of the key-value pairs in the Data table are strings unless otherwise specified. Codes with the CONFIG prefix have to do with the configuration of the corpora associated with the build. Codes with the MODEL prefix have to do with constraints imposed by the model (e.g., NLLB). Codes with the USFM prefix have to do with problems with the USFM content of a project selected for training or inferencing. A maximum of 1000 diagnostics can be emitted by a build. If more than 1000 are generated, only the first 1000 (sorted high severity to low severity) are emitted and DiagnosticsTruncated will be set to true.

Code CONFIG-0001

Description: One of the Paratext projects referenced in this build’s corpora is the daughter project of a project which is not among the projects referenced in this build’s corpora. The daughter project will default to being extracted with English versification.

Message: Unable to locate parent project parentProjectName parentProjectGuid of daughter project daughterProjectName daughterProjectGuid (parallel corpus parallelCorpusId, monolingual corpus monolingualCorpusId).

Severity: Warn

Data:

Key	Value
parentProjectName	The short name of the parent Paratext project which could not be located.
parentProjectGuid	The GUID of the parent Paratext project which could not be located.
daughterProjectName	The short name of the daughter Paratext project whose parent could not be located.
daughterProjectGuid	The GUID of the daughter Paratext project whose parent could not be located.
parallelCorpusId	The id of the parallel corpus in which this was detected.
monolingualCorpusId	The id of the monolingual corpus in which this was detected.

Code CONFIG-0002

Description: Among Paratext projects selected for training or inferencing, more than one versification is represented. In some circumstances, Serval will not be able to align verses across these versifications properly.

Message: There are multiple versifications represented among Paratext projects selected for training or inferencing: projectVersifications.

Severity: Info

Data:

Key	Value
projectVersifications (Dictionarystring, string)	A map of versification names (e.g. Original) per Paratext project GUID

Code CONFIG-0003

Description: There are fewer than 600 (but more than 0) segments selected for training; this is less than recommended.

Message: Only trainCount segments were selected for training. Training on fewer than minimumTrainCount segments is not recommended.

Severity: Warn

Data:

Key	Value
trainCount (int)	The number of parallel segments selected for training.
minimumTrainCount (int)	The minimum recommended number of parallel segments selected for training

Code CONFIG-0004

Description: There was no data specified for inferencing and the model is not persisted.

Message: There was no data specified for inferencing and the model is not persisted.

Severity: Error

Data: [ ]

Code MODEL-0001

Description: The resolved script for the engine source language is not known to the base model.

Message: The script for the source language ‘resolvedCode’ is not known to the base model modelName.

Severity: Warn

Data:

Key	Value
resolvedCode	The language code that the engine language has resolved to (e.g., spa_Latn).
modelName	The name of the base model (e.g. NLLB).

Code MODEL-0002

Description: The resolved script for the engine target language is not known to the base model.

Message: The script for the target language ‘resolvedCode’ is not known to the base model modelName.

Severity: Warn

Data:

Key	Value
resolvedCode	The language code that the engine language has resolved to (e.g., spa_Latn).
modelName	The name of the base model (e.g. NLLB).

Code MODEL-0003

Description: The average pretranslation model confidence for this book is unusually low, so the quality of the draft associated with it is likely poor. This may indicate a bad training set up.

Message: The average pretranslation model confidence averagePretranslationConfidence in book bookId is unusually low for the base model modelName.

Severity: Warn

Data:

Key	Value
averagePretranslationConfidence (double)	The language code that the engine language has resolved to (e.g., spa_Latn).
bookId	The three letter book ID for the Scripture book (e.g. MAT).
modelName	The name of the base model (e.g. NLLB).

Code MODEL-0004

Description: At least one of the language codes associated with this build is unknown to the base model and there is no training data selected.

Message: The following language codes are unknown to the base model modelName: unknownLanguageCodes; and no language data was selected for training.

Severity: Error

Data:

Key	Value
modelName	The name of the base model (e.g. NLLB).
unknownLanguageCodes (Liststring)	The language codes that are unknown to the model (e.g., spa_Latn).

Code USFM-0001

Description: One of the Paratext projects referenced in this build’s corpora contains USFM with an invalid chapter number (e.g. \c $). Serval may not be able to extract any of the segments in this chapter properly.

Message: Invalid chapter number in project projectName projectGuid at usfmFileName line lineNumber, verse verseReference (parallel corpus parallelCorpusId, monolingual corpus monolingualCorpusId).

Severity: Warn

Data:

Key	Value
projectName	The short name of the Paratext project in which this was detected.
projectGuid	The GUID of the Paratext project in which this was detected.
usfmFileName	The name of the USFM file in which this was detected.
lineNumber (int)	The line number at which this was detected.
verseReference	The verse reference at which this was detected.
parallelCorpusId	The id of the parallel corpus in which this was detected.
monolingualCorpusId	The id of the monolingual corpus in which this was detected.

Code USFM-0002

Description: One of the Paratext projects referenced in this build’s corpora contains USFM with an invalid verse number (e.g. \v $). Serval will not be able to align this verse segment properly.

Message: Invalid verse number in project projectName projectGuid at usfmFileName line lineNumber, verse verseReference (parallel corpus parallelCorpusId, monolingual corpus monolingualCorpusId).

Severity: Warn

Data:

Key	Value
projectName	The short name of the Paratext project in which this was detected.
projectGuid	The GUID of the Paratext project in which this was detected.
usfmFileName	The name of the USFM file in which this was detected.
lineNumber (int)	The line number at which this was detected.
verseReference	The verse reference at which this was detected.
parallelCorpusId	The id of the parallel corpus in which this was detected.
monolingualCorpusId	The id of the monolingual corpus in which this was detected.

Code USFM-0003

Description: One of the Paratext projects referenced in this build’s corpora contains USFM with additional verses which were selected for training or inferencing but are not specified in the project’s versification. Serval may not be able to align these verses properly.

Message: numberOfVerses extra verses in project projectName projectGuid at usfmFileName lines lineNumbers, verses verseReferences (parallel corpus parallelCorpusId, monolingual corpus monolingualCorpusId).

Severity: Info

Data:

Key	Value
numberOfVerses (int)	The number of affected verses.
projectName	The short name of the Paratext project in which this was detected.
projectGuid	The GUID of the Paratext project in which this was detected.
usfmFileName	The name of the USFM file in which this was detected.
lineNumber (Listint)	The line numbers at which this was detected.
verseReference (Liststring)	The verse references at which this was detected.
parallelCorpusId	The id of the parallel corpus in which this was detected.
monolingualCorpusId	The id of the monolingual corpus in which this was detected.

Code USFM-0004

Description: One of the Paratext projects referenced in this build’s corpora contains USFM that is missing verses which were selected for training or inferencing.

Message: Missing numberOfVerses verses in project projectName projectGuid at usfmFileName lines lineNumbers, verse verseReferences (parallel corpus parallelCorpusId, monolingual corpus monolingualCorpusId).

Severity: Warn

Data:

Key	Value
numberOfVerses (int)	The number of affected verses.
projectName	The short name of the Paratext project in which this was detected.
projectGuid	The GUID of the Paratext project in which this was detected.
usfmFileName	The name of the USFM file in which this was detected.
lineNumber (Listint)	The line numbers at which this was detected.
verseReference (Liststring)	The verse references at which this was detected.
parallelCorpusId	The id of the parallel corpus in which this was detected.
monolingualCorpusId	The id of the monolingual corpus in which this was detected.

Code USFM-0005

Description: One of the Paratext projects referenced in this build’s corpora contains USFM with verses which were selected for training or inferencing whose verse segments are not consistent with the project’s versification.

Message: Incorrect verse segment in project projectName projectGuid at usfmFileName line lineNumber, verse verseReference (parallel corpus parallelCorpusId, monolingual corpus monolingualCorpusId).

Severity: Info

Data:

Key	Value
projectName	The short name of the Paratext project in which this was detected.
projectGuid	The GUID of the Paratext project in which this was detected.
usfmFileName	The name of the USFM file in which this was detected.
lineNumber (int)	The line number at which this was detected.
verseReference	The verse reference at which this was detected.
parallelCorpusId	The id of the parallel corpus in which this was detected.
monolingualCorpusId	The id of the monolingual corpus in which this was detected.

Code USFM-0006

Description: One of the Paratext projects referenced in this build’s corpora contains USFM with a verse range selected for training or inferencing that would cross a chapter boundary if converted to Original versification. Serval may not be able to align these verses properly if training or inferencing across versifications.

Message: Unsupported verse range in project projectName projectGuid at usfmFileName line lineNumber, verse verseReference (parallel corpus parallelCorpusId, monolingual corpus monolingualCorpusId).

Severity: Info

Data:

Key	Value
projectName	The short name of the Paratext project in which this was detected.
projectGuid	The GUID of the Paratext project in which this was detected.
usfmFileName	The name of the USFM file in which this was detected.
lineNumber (int)	The line number at which this was detected.
verseReference	The verse reference at which this was detected.
parallelCorpusId	The id of the parallel corpus in which this was detected.
monolingualCorpusId	The id of the monolingual corpus in which this was detected.