fix: 数据库修改日志目录权限/修复图片相似度重复重试任务时误过滤

deployment/helm/datamate/charts/database/templates/deployment.yaml

@@ -32,6 +32,15 @@ spec:
       securityContext:
         {{- toYaml . | nindent 8 }}
       {{- end }}
+      {{- with .Values.initContainers }}
+      initContainers:
+        {{- $root := $ }}
+        {{- range . }}
+        {{- $defaultImage := dict "image" (include "database.image" $root) }}
+        {{- $containerConfig := merge (deepCopy .) $defaultImage }}
+        - {{ toYaml $containerConfig | nindent 10 | trim }}
+        {{- end }}
+      {{- end }}
       containers:
         - name: {{ .Chart.Name }}
           {{- with .Values.securityContext }}

deployment/helm/datamate/charts/database/values.yaml

@@ -106,3 +106,5 @@ nodeSelector: {}
 tolerations: []
 
 affinity: {}
+
+initContainers: []

deployment/helm/datamate/charts/frontend/templates/deployment.yaml

@@ -14,7 +14,6 @@ spec:
   template:
     metadata:
       annotations:
-        checksum/nginx: {{ include (print .Template.BasePath "/configmap.yaml") . | sha256sum }}
         {{- with .Values.podAnnotations }}
         {{- toYaml . | nindent 8 }}
         {{- end }}

deployment/helm/datamate/values.yaml

@@ -68,6 +68,19 @@ operatorVolume: &operatorVolume
     claimName: datamate-operator-pvc
 
 database:
+  initContainers:
+    - name: init-chown
+      imagePullPolicy: IfNotPresent
+      command:
+        - sh
+        - -c
+        - "chown -R 999:999 /var/log/datamate/database && chmod 750 /var/log/datamate/database"
+      securityContext:
+        runAsUser: 0
+      volumeMounts:
+        - name: log-volume
+          mountPath: /var/log/datamate/database
+          subPath: database
   env:
     - name: POSTGRES_USER
       value: postgres

runtime/ops/filter/img_similar_images_cleaner/process.py

@@ -49,7 +49,6 @@ class ImgSimilarImagesCleaner(Filter):
     """去除相似图片的插件"""
 
     DEFAULT_SIMILAR_THRESHOLD = 0.8  # 默认相似度阈值
-    DEFAULT_TASK_UUID = "uuid"  # 默认任务UUID
     DEFAULT_ORB_RATIO = 0.8  # 默认特征点距离比率
     DEFAULT_MIX_SIMILARITY = 0.75  # 默认相似度算法阈值
     DEFAULT_IMG_RESIZE = 200  # 默认图片压缩尺寸
@@ -59,7 +58,7 @@ def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
         self.similar_threshold = kwargs.get("similarThreshold", self.DEFAULT_SIMILAR_THRESHOLD)  # 默认相似度阈值为0.8
         # task_uuid为标识该数据集的唯一标志
-        self.task_uuid = kwargs.get("uuid", self.DEFAULT_TASK_UUID)
+        self.task_uuid = kwargs.get("uuid", "")
         self.orb_ratio = self.DEFAULT_ORB_RATIO  # 特征点距离的比率，该数值为经验值
         self.mix_similarity = self.DEFAULT_MIX_SIMILARITY  # 选择相似度算法的阈值，该数值为经验值
         self.img_resize = self.DEFAULT_IMG_RESIZE  # 图片压缩尺寸
@@ -230,6 +229,7 @@ def execute(self, sample: Dict[str, Any]) -> Dict[str, Any]:
         self.read_file_first(sample)
         file_name = sample[self.filename_key]
         img_bytes = sample[self.data_key]
+        self.task_uuid = sample.get("instance_id") if not self.task_uuid else self.task_uuid
         data = bytes_to_numpy(img_bytes) if img_bytes else np.array([])
         similar_images = self.filter_similar_images(data, file_name)
         # 若相似图片，sample[self.data_key]设为空