init datamate

backend/services/data-management-service/src/main/resources/config/application-datamanagement.yml

@@ -0,0 +1,11 @@
+dataMate:
+  datamanagement:
+    file-storage:
+      upload-dir: ${FILE_UPLOAD_DIR:./uploads}
+      max-file-size: 10485760   # 10MB
+      max-request-size: 52428800 # 50MB
+    cache:
+      ttl: 3600
+      max-size: 1000
+# MyBatis is configured centrally in main-application (mapper-locations & aliases)
+# to avoid list overriding issues when importing multiple module configs.

backend/services/data-management-service/src/main/resources/mappers/DatasetFileMapper.xml

@@ -0,0 +1,98 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE mapper PUBLIC "-//mybatis.org//DTD Mapper 3.0//EN"
+    "http://mybatis.org/dtd/mybatis-3-mapper.dtd">
+<mapper namespace="com.datamate.datamanagement.infrastructure.persistence.mapper.DatasetFileMapper">
+    <sql id="Base_Column_List">
+      id, dataset_id, file_name, file_path, file_type, file_size, check_sum, tags, metadata, status,
+          upload_time, last_access_time, created_at, updated_at
+    </sql>
+
+  <select id="findById" parameterType="string"
+          resultType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_dataset_files
+        WHERE id = #{id}
+    </select>
+
+  <select id="findByDatasetId" parameterType="string"
+          resultType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_dataset_files
+        WHERE dataset_id = #{datasetId}
+        ORDER BY upload_time DESC
+    </select>
+
+  <select id="findByDatasetIdAndStatus" resultType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_dataset_files
+        WHERE dataset_id = #{datasetId}
+          AND status = #{status}
+        ORDER BY upload_time DESC
+    </select>
+
+  <select id="findByDatasetIdAndFileType" resultType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_dataset_files
+        WHERE dataset_id = #{datasetId}
+          AND file_type = #{fileType}
+        ORDER BY upload_time DESC
+    </select>
+
+    <select id="countByDatasetId" parameterType="string" resultType="long">
+        SELECT COUNT(*) FROM t_dm_dataset_files WHERE dataset_id = #{datasetId}
+    </select>
+
+    <select id="countCompletedByDatasetId" parameterType="string" resultType="long">
+        SELECT COUNT(*) FROM t_dm_dataset_files WHERE dataset_id = #{datasetId} AND status = 'COMPLETED'
+    </select>
+
+    <select id="sumSizeByDatasetId" parameterType="string" resultType="long">
+        SELECT COALESCE(SUM(file_size), 0) FROM t_dm_dataset_files WHERE dataset_id = #{datasetId}
+    </select>
+
+  <select id="findByDatasetIdAndFileName" resultType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_dataset_files
+        WHERE dataset_id = #{datasetId} AND file_name = #{fileName}
+        LIMIT 1
+    </select>
+
+  <select id="findAllByDatasetId" parameterType="string"
+          resultType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_dataset_files
+        WHERE dataset_id = #{datasetId}
+        ORDER BY upload_time DESC
+    </select>
+
+  <select id="findByCriteria" resultType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_dataset_files
+        WHERE dataset_id = #{datasetId}
+        <!-- Replace invalid XML '&&' with 'and' for MyBatis OGNL -->
+        <if test="fileType != null and fileType != ''">
+            AND file_type = #{fileType}
+        </if>
+        <if test="status != null and status != ''">
+            AND status = #{status}
+        </if>
+        ORDER BY upload_time DESC
+    </select>
+
+
+    <update id="update" parameterType="com.datamate.datamanagement.domain.model.dataset.DatasetFile">
+        UPDATE t_dm_dataset_files
+        SET file_name = #{fileName},
+            file_path = #{filePath},
+            file_type = #{fileType},
+            file_size = #{fileSize},
+            upload_time = #{uploadTime},
+            last_access_time = #{lastAccessTime},
+            status = #{status}
+        WHERE id = #{id}
+    </update>
+
+    <delete id="deleteById" parameterType="string">
+        DELETE FROM t_dm_dataset_files WHERE id = #{id}
+    </delete>
+</mapper>

backend/services/data-management-service/src/main/resources/mappers/DatasetMapper.xml

@@ -0,0 +1,152 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE mapper PUBLIC "-//mybatis.org//DTD Mapper 3.0//EN"
+    "http://mybatis.org/dtd/mybatis-3-mapper.dtd">
+<mapper namespace="com.datamate.datamanagement.infrastructure.persistence.mapper.DatasetMapper">
+
+  <sql id="Base_Column_List">
+    id, name, description, dataset_type, category, path, format, schema_info, size_bytes, file_count, record_count,
+        retention_days, tags, metadata, status, is_public, is_featured, version, created_at, updated_at, created_by, updated_by
+  </sql>
+
+  <sql id="Alias_D_Column_List">
+        d.id AS id,
+        d.name AS name,
+        d.description AS description,
+        d.dataset_type AS dataset_type,
+        d.category AS category,
+        d.path AS path,
+        d.format AS format,
+        d.schema_info AS schema_info,
+        d.size_bytes AS size_bytes,
+        d.file_count AS file_count,
+        d.record_count AS record_count,
+        d.retention_days AS retention_days,
+        d.tags AS tags,
+        d.metadata AS metadata,
+        d.status AS status,
+        d.is_public AS is_public,
+        d.is_featured AS is_featured,
+        d.version AS version,
+        d.created_at AS created_at,
+        d.updated_at AS updated_at,
+        d.created_by AS created_by,
+        d.updated_by AS updated_by
+  </sql>
+
+  <select id="findById" parameterType="string" resultType="com.datamate.datamanagement.domain.model.dataset.Dataset">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_datasets
+        WHERE id = #{id}
+    </select>
+
+  <select id="findByName" parameterType="string"
+          resultType="com.datamate.datamanagement.domain.model.dataset.Dataset">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_datasets
+        WHERE name = #{name}
+        LIMIT 1
+    </select>
+
+  <select id="findByStatus" parameterType="string"
+          resultType="com.datamate.datamanagement.domain.model.dataset.Dataset">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_datasets
+        WHERE status = #{status}
+        ORDER BY updated_at DESC
+    </select>
+
+  <select id="findByCreatedBy" resultType="com.datamate.datamanagement.domain.model.dataset.Dataset">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_datasets
+        WHERE created_by = #{createdBy}
+        ORDER BY created_at DESC
+    </select>
+
+  <select id="findByTypeCode" resultType="com.datamate.datamanagement.domain.model.dataset.Dataset">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_datasets
+        WHERE dataset_type = #{typeCode}
+        ORDER BY created_at DESC
+    </select>
+
+  <select id="findByTagNames" resultType="com.datamate.datamanagement.domain.model.dataset.Dataset">
+        SELECT DISTINCT <include refid="Alias_D_Column_List"/>
+        FROM t_dm_datasets d
+        JOIN t_dm_dataset_tags dt ON d.id = dt.dataset_id
+        JOIN t_dm_tags t ON t.id = dt.tag_id
+        WHERE t.name IN
+        <foreach collection="tagNames" item="name" open="(" separator="," close=")">
+            #{name}
+        </foreach>
+        ORDER BY d.created_at DESC
+    </select>
+
+  <select id="findByKeyword" resultType="com.datamate.datamanagement.domain.model.dataset.Dataset">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_datasets
+        WHERE name LIKE CONCAT('%', #{keyword}, '%')
+           OR description LIKE CONCAT('%', #{keyword}, '%')
+        ORDER BY created_at DESC
+    </select>
+
+  <select id="findByCriteria" resultType="com.datamate.datamanagement.domain.model.dataset.Dataset">
+        SELECT DISTINCT <include refid="Alias_D_Column_List"/>
+        FROM t_dm_datasets d
+        LEFT JOIN t_dm_dataset_tags dt ON d.id = dt.dataset_id
+        LEFT JOIN t_dm_tags t ON t.id = dt.tag_id
+        <where>
+            <if test="typeCode != null and typeCode != ''">
+                AND d.dataset_type = #{typeCode}
+            </if>
+            <if test="status != null and status != ''">
+                AND d.status = #{status}
+            </if>
+            <if test="keyword != null and keyword != ''">
+                AND (d.name LIKE CONCAT('%', #{keyword}, '%') OR d.description LIKE CONCAT('%', #{keyword}, '%'))
+            </if>
+            <if test="tagNames != null and tagNames.size > 0">
+                AND t.name IN
+                <foreach collection="tagNames" item="n" open="(" separator="," close=")">
+                    #{n}
+                </foreach>
+            </if>
+        </where>
+        ORDER BY d.created_at DESC
+    </select>
+
+    <select id="countByCriteria" resultType="long">
+        SELECT COUNT(DISTINCT d.id)
+        FROM t_dm_datasets d
+        LEFT JOIN t_dm_dataset_tags dt ON d.id = dt.dataset_id
+        LEFT JOIN t_dm_tags t ON t.id = dt.tag_id
+        <where>
+            <if test="typeCode != null and typeCode != ''">
+                AND d.dataset_type = #{typeCode}
+            </if>
+            <if test="status != null and status != ''">
+                AND d.status = #{status}
+            </if>
+            <if test="keyword != null and keyword != ''">
+                AND (d.name LIKE CONCAT('%', #{keyword}, '%') OR d.description LIKE CONCAT('%', #{keyword}, '%'))
+            </if>
+            <if test="tagNames != null and tagNames.size > 0">
+                AND t.name IN
+                <foreach collection="tagNames" item="n" open="(" separator="," close=")">
+                    #{n}
+                </foreach>
+            </if>
+        </where>
+    </select>
+
+    <delete id="deleteById" parameterType="string">
+        DELETE FROM t_dm_datasets WHERE id = #{id}
+    </delete>
+
+    <select id="getAllDatasetStatistics" resultType="com.datamate.datamanagement.interfaces.dto.AllDatasetStatisticsResponse">
+      SELECT
+        COUNT(*) AS total_datasets,
+        SUM(size_bytes) AS total_size,
+        SUM(file_count) AS total_files
+      FROM t_dm_datasets;
+    </select>
+</mapper>

backend/services/data-management-service/src/main/resources/mappers/TagMapper.xml

@@ -0,0 +1,111 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE mapper PUBLIC "-//mybatis.org//DTD Mapper 3.0//EN"
+    "http://mybatis.org/dtd/mybatis-3-mapper.dtd">
+<mapper namespace="com.datamate.datamanagement.infrastructure.persistence.mapper.TagMapper">
+
+    <resultMap id="TagResultMap" type="com.datamate.datamanagement.domain.model.dataset.Tag">
+        <id column="id" property="id"/>
+        <result column="name" property="name"/>
+        <result column="description" property="description"/>
+        <result column="category" property="category"/>
+        <result column="color" property="color"/>
+        <result column="usage_count" property="usageCount"/>
+        <result column="created_at" property="createdAt"/>
+        <result column="updated_at" property="updatedAt"/>
+    </resultMap>
+
+    <sql id="Base_Column_List">
+        id, name, description, category, color, usage_count, created_at, updated_at
+    </sql>
+
+    <select id="findById" parameterType="string" resultMap="TagResultMap">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_tags
+        WHERE id = #{id}
+    </select>
+
+    <select id="findByName" parameterType="string" resultMap="TagResultMap">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_tags
+        WHERE name = #{name}
+        LIMIT 1
+    </select>
+
+    <select id="findByNameIn" parameterType="list" resultMap="TagResultMap">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_tags
+        WHERE name IN
+        <foreach collection="list" item="n" open="(" separator="," close=")">
+            #{n}
+        </foreach>
+    </select>
+
+    <select id="findByKeyword" parameterType="string" resultMap="TagResultMap">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_tags
+        WHERE name LIKE CONCAT('%', #{keyword}, '%')
+        ORDER BY usage_count DESC, name ASC
+    </select>
+
+    <select id="findAllByOrderByUsageCountDesc" resultMap="TagResultMap">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_tags
+        ORDER BY usage_count DESC, name ASC
+    </select>
+
+    <insert id="insert" parameterType="com.datamate.datamanagement.domain.model.dataset.Tag">
+        INSERT INTO t_dm_tags (id, name, description, category, color, usage_count)
+        VALUES (#{id}, #{name}, #{description}, #{category}, #{color}, #{usageCount})
+    </insert>
+
+    <update id="update" parameterType="com.datamate.datamanagement.domain.model.dataset.Tag">
+        UPDATE t_dm_tags
+        SET name = #{name},
+            description = #{description},
+            category = #{category},
+            color = #{color},
+            usage_count = #{usageCount}
+        WHERE id = #{id}
+    </update>
+
+    <update id="updateUsageCount">
+        UPDATE t_dm_tags
+        SET usage_count = #{usageCount}
+        WHERE id = #{id}
+    </update>
+
+    <!-- Dataset & Tag relations -->
+    <insert id="insertDatasetTag">
+        INSERT INTO t_dm_dataset_tags (dataset_id, tag_id)
+        VALUES (#{datasetId}, #{tagId})
+    </insert>
+
+    <delete id="deleteDatasetTagsByDatasetId">
+        DELETE FROM t_dm_dataset_tags WHERE dataset_id = #{datasetId}
+    </delete>
+
+    <select id="findByDatasetId" parameterType="string" resultMap="TagResultMap">
+        SELECT t.id, t.name, t.description, t.category, t.color, t.usage_count, t.created_at, t.updated_at
+        FROM t_dm_tags t
+        JOIN t_dm_dataset_tags dt ON dt.tag_id = t.id
+        WHERE dt.dataset_id = #{datasetId}
+        ORDER BY t.usage_count DESC, t.name ASC
+    </select>
+
+    <delete id="deleteTagsById">
+        DELETE FROM t_dm_tags WHERE
+        id IN
+        <foreach collection="ids" item="id" open="(" separator="," close=")">
+          #{id}
+        </foreach>
+    </delete>
+
+    <select id="findByIdIn" resultMap="TagResultMap">
+        SELECT <include refid="Base_Column_List"/>
+        FROM t_dm_tags
+        WHERE id IN
+        <foreach collection="ids" item="id" open="(" separator="," close=")">
+          #{id}
+        </foreach>
+    </select>
+</mapper>