## Status

- ACCEPTED on 2024-02-07

## Context

The Gradle code base is essentially a large monolith, without strong internal boundaries.
This has a number of negative effects on productivity, including:

- Unclear ownership of code.
- Difficult to focus on one particular area.
- Unintended coupling between areas of the code, including tests.

## Decision

        }
        return intentDetectionPrompt;
    }

    /**
     * Gets the system prompt for unclear intent responses.
     *
     * @return the unclear intent system prompt
     */
    protected String getUnclearIntentSystemPrompt() {
        if (unclearIntentSystemPrompt == null) {

   * have started execution. This applies also to tasks given to {@code invokeAll} or {@code
   * invokeAny} which are pending serial execution, even the subset of the tasks that have not yet
   * started execution. It is unclear from the {@code ExecutorService} specification if these should
   * be included, and it's much easier to implement the interpretation that they not be. Finally, a

            LlmStreamCallback callback);

    /**
     * Generates a response asking user for clarification when intent is unclear.
     *
     * @param userMessage the user's message
     * @param history the conversation history
     * @param callback the streaming callback
     */

                logger.debug("[RAG] Intent detected. intent={}, query={}", intentResult.getIntent(), intentResult.getQuery());
            }

            if (intentResult.getIntent() == ChatIntent.UNCLEAR) {
                // Unclear intent - generate answer with empty documents to ask for clarification
                final LlmChatResponse llmResponse = llmClientManager.generateAnswer(userMessage, Collections.emptyList(), history);

 *     "Unexpected bill status: %s", bill.status());
 * }
 *
 * <h3>Comparison to alternatives</h3>
 *
 * <p><b>Note:</b> In some cases the differences explained below can be subtle. When it's unclear
 * which approach to use, <b>don't worry</b> too much about it; just pick something that seems
 * reasonable and it will be fine.
 *
 * <ul>

repeats of these filename(s) with different casing:
EOF
    find . | tr '[A-Z]' '[a-z]' | sort | uniq -d | tee $BATS_FILE_TMPDIR/repeats
    [[ ! -s $BATS_FILE_TMPDIR/repeats ]]
}

# It's unclear why, but running this on //tensorflow/... is faster than running
# only on affected targets, usually. There are targets in //tensorflow/lite that
# don't pass --nobuild, so they're on their own.
#

         * annotations. This is intentional. The semantics of annotation handling during
         * type resolution are unclear and have changed across Java versions. Until there's
         * a clear specification for what annotations should mean on resolved TypeVariables
         * with modified bounds, annotation methods will throw

 *     "Unexpected bill status: %s", bill.status());
 * }
 *
 * <h3>Comparison to alternatives</h3>
 *
 * <p><b>Note:</b> In some cases the differences explained below can be subtle. When it's unclear
 * which approach to use, <b>don't worry</b> too much about it; just pick something that seems
 * reasonable and it will be fine.
 *
 * <ul>

   * have started execution. This applies also to tasks given to {@code invokeAll} or {@code
   * invokeAny} which are pending serial execution, even the subset of the tasks that have not yet
   * started execution. It is unclear from the {@code ExecutorService} specification if these should
   * be included, and it's much easier to implement the interpretation that they not be. Finally, a