chore: migrate java-firestore to google-cloud-java

[suztomo]

b/477663340

[gemini-code-assist]

Code Review

This pull request introduces the Java Firestore SDK, including the core library and admin client, along with Kokoro configurations, documentation, and metadata. The implementation supports features like bulk writes, aggregate queries, and data pipelines. Feedback focuses on several potential NullPointerException vulnerabilities in FirestoreBundle, DocumentSnapshot, and PipelineResult when processing nullable timestamps or non-document results. Additionally, a security concern was raised regarding the exposure of sensitive tokens in a script's git configuration, and improvements were suggested for error handling in scripts and the replacement of deprecated Double constructors.

[gemini-code-assist]

Including the ${GH_TOKEN} directly in the remote URL is a security risk as it will be stored in plain text in the .git/config file. It is safer to use the token only during the push command or configure it via a header to avoid persistence in the local configuration.

[gemini-code-assist]

Potential NullPointerException when calling documentSnapshot.getReadTime().compareTo(...) because getReadTime() is @Nullable. The code should use the snapReadTime variable calculated earlier (on line 89), which correctly handles the null case by defaulting to Timestamp.MIN_VALUE.

      if (snapReadTime.compareTo(
              Timestamp.fromProto(originalDocument.getMetadata().getReadTime()))
           > 0) {
        latestReadTime = snapReadTime;
      }

[gemini-code-assist]

Potential NullPointerException in toDocumentPb() if createTime or updateTime are null. These fields are @Nullable and can be null if the snapshot was created from an object (e.g., via fromObject).

    if (createTime != null) {
      document.setCreateTime(createTime.toProto());
    }
    if (updateTime != null) {
      document.setUpdateTime(updateTime.toProto());
    }

[gemini-code-assist]

Potential NullPointerException in getId() if docRef is null. The Javadoc states that it returns null if the result does not correspond to a document, but the current implementation will throw an exception.

  @Nullable
  @BetaApi
  public String getId() {
    return docRef != null ? docRef.getId() : null;
  }

[gemini-code-assist]

Potential NullPointerException in toDocumentPb() if docRef, createTime, or updateTime are null. These fields can be null for non-document results or results without timestamps (like those from aggregations).

    if (docRef != null) {
      document.setName(docRef.getName());
    }
    document.putAllFields(fields);
    if (createTime != null) {
      document.setCreateTime(createTime.toProto());
    }
    if (updateTime != null) {
      document.setUpdateTime(updateTime.toProto());
    }

[gemini-code-assist]

The error message uses an undefined variable ${json_content} and incorrectly refers to the response as JSON, while the URL on line 19 points to an XML file (maven-metadata.xml). It should use ${xml_content} instead.

Suggested change

	echo "The latest version of ${group_id}:${artifact_id} is empty."
	echo "The returned json from maven.org is invalid: ${json_content}"
	echo "The latest version of ${group_id}:${artifact_id} is empty."
	echo "The returned XML from maven.org is invalid: ${xml_content}"
	exit 1

[gemini-code-assist]

Use of deprecated new Double(double) constructor. It is recommended to use Double.valueOf(double) instead for better performance and future compatibility.

Suggested change

	* @param initialOpsPerSecond The initial maximum number of operations per second allowed by the
	return setInitialOpsPerSecond(Double.valueOf(initialOpsPerSecond));

[gemini-code-assist]

Use of deprecated new Double(double) constructor. It is recommended to use Double.valueOf(double) instead.

Suggested change

	* @param maxOpsPerSecond The maximum number of operations per second allowed by the throttler.
	return setMaxOpsPerSecond(Double.valueOf(maxOpsPerSecond));