Update main.yml (#3391)

ansible/roles/ml-analytics-service/defaults/main.yml

@@ -51,10 +51,10 @@ ml_analytics_api_access_token: "{{ml_api_access_token | default('ml_core_interna
 ml_analytics_druid_observation_status_injestion_spec : '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type": "azure","uris": ["azure://telemetry-data-store/observation/status/sl_observation_status.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"sl-observation-status","granularitySpec":{"type":"uniform","queryGranularity":"DAY","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"updatedAt","format":"auto"},"dimensionsSpec":{"dimensions":["status","entity_externalId","entity_id","entity_type","solution_id","solution_externalId","submission_id","entity_name","solution_name","role_title","school_name","school_id","school_externalId","state_name","state_id","state_externalId","district_name","district_id","district_externalId","zone_name","zone_id","zone_externalId","parent_name","parent_id","block_name","block_id","block_externalId","cluster_name","cluster_id","cluster_externalId","teacher_name","teacher_id","teacher_externalId","schoolLeader_name","schoolLeader_id","schoolLeader_externalId","taluk_id","taluk_externalId","taluk_name","hub_name","hub_id","hub_externalId","mandal_name","mandal_id","mandal_externalId","region_name","region_id","region_externalId","complex_name","complex_id","complex_externalId","completedDate","beat_name","beat_id","beat_externalId","division_name","division_id","division_externalId","Village/Ward_name","Village/Ward_id","Village/Ward_externalId","channel","parent_channel","program_id","program_externalId","program_name","app_name","user_id","private_program","solution_type","organisation_name", "ecm_marked_na","board_name","updatedAt","organisation_id"]},"metricsSpec":[]}}}'
 ml_analytics_druid_project_injestion_spec : '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type": "azure","uris": ["azure://telemetry-data-store/projects/sl_projects.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"sl-project","granularitySpec":{"type":"uniform","queryGranularity":"DAY","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"project_updated_date","format":"auto"},"dimensionsSpec":{"dimensions":[]},"metricsSpec":[]}}}'
 ml_analytics_azure_sas_token: "{{ sunbird_private_storage_account_key }}"
-ml_analytics_druid_distinctCnt_obs_injestion_spec: '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type":"azure","uris":["azure://telemetry-data-store/observation/distinctCount/ml_observation_distinctCount_status.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"ml-distinctCnt-obs-status","granularitySpec":{"type":"uniform","queryGranularity":"none","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"time_stamp","format":"auto"},"dimensionsSpec":{"dimensions":[{"type":"string","name":"program_name"},{"type":"string","name":"program_id"},{"type":"string","name":"solution_name"},{"type":"string","name":"solution_id"},{"type":"string","name":"status"},{"type":"string","name":"state_name"},{"type":"string","name":"state_externalId"},{"type":"string","name":"district_name"},{"type":"string","name":"district_externalId"},{"type":"string","name":"organisation_name"},{"type":"string","name":"organisation_id"},{"type":"string","name":"parent_channel"},{"type":"string","name":"solution_type"},{"type":"string","name":"private_program"},{"type":"long","name":"unique_submissions"},{"type":"long","name":"unique_entities"},{"type":"long","name":"unique_users"},{"type":"string","name":"time_stamp"}]},"metricsSpec":[]}}}'
-ml_analytics_druid_distinctCnt_obs_domain_injestion_spec: '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type":"azure","uris":["azure://telemetry-data-store/observation/distinctCount_domain/ml_observation_distinctCount_domain.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"ml-distinctCnt-obs-domain","granularitySpec":{"type":"uniform","queryGranularity":"none","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"time_stamp","format":"auto"},"dimensionsSpec":{"dimensions":[{"type":"string","name":"program_name"},{"type":"string","name":"program_id"},{"type":"string","name":"solution_name"},{"type":"string","name":"solution_id"},{"type":"string","name":"state_name"},{"type":"string","name":"state_externalId"},{"type":"string","name":"district_name"},{"type":"string","name":"district_externalId"},{"type":"string","name":"organisation_name"},{"type":"string","name":"organisation_id"},{"type":"string","name":"parent_channel"},{"type":"string","name":"solution_type"},{"type":"string","name":"private_program"},{"type":"string","name":"domain_name"},{"type":"string","name":"domain_externalId"},{"type":"string","name":"domain_level"},{"type":"long","name":"unique_submissions"},{"type":"long","name":"unique_entities"},{"type":"long","name":"unique_users"},{"type":"string","name":"time_stamp"}]},"metricsSpec":[]}}}'
-ml_analytics_druid_distinctCnt_obs_domain_criteria_injestion_spec: '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type":"azure","uris":["azure://telemetry-data-store/observation/distinctCount_domain_criteria/ml_observation_distinctCount_domain_criteria.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"ml-distinctCnt-obs-domain-criteria","granularitySpec":{"type":"uniform","queryGranularity":"none","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"time_stamp","format":"auto"},"dimensionsSpec":{"dimensions":[{"type":"string","name":"program_name"},{"type":"string","name":"program_id"},{"type":"string","name":"solution_name"},{"type":"string","name":"solution_id"},{"type":"string","name":"state_name"},{"type":"string","name":"state_externalId"},{"type":"string","name":"district_name"},{"type":"string","name":"district_externalId"},{"type":"string","name":"organisation_name"},{"type":"string","name":"organisation_id"},{"type":"string","name":"parent_channel"},{"type":"string","name":"solution_type"},{"type":"string","name":"private_program"},{"type":"string","name":"domain_name"},{"type":"string","name":"domain_externalId"},{"type":"string","name":"domain_level"},{"type":"string","name":"criteria_name"},{"type":"string","name":"criteria_score"},{"type":"string","name":"criteria_id"},{"type":"long","name":"unique_submissions"},{"type":"long","name":"unique_entities"},{"type":"long","name":"unique_users"},{"type":"string","name":"time_stamp"}]},"metricsSpec":[]}}}'
-ml_analytics_druid_distinctCnt_projects_status_injestion_spec: '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type":"azure","uris":["azure://telemetry-data-store/projects/distinctCount/ml_projects_distinctCount.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"ml-distinctCnt-projects-status","granularitySpec":{"type":"uniform","queryGranularity":"none","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"time_stamp","format":"auto"},"dimensionsSpec":{"dimensions":[{"type":"string","name":"program_name"},{"type":"string","name":"program_id"},{"type":"string","name":"project_title"},{"type":"string","name":"solution_id"},{"type":"string","name":"status_of_project"},{"type":"string","name":"state_name"},{"type":"string","name":"state_externalId"},{"type":"string","name":"district_name"},{"type":"string","name":"district_externalId"},{"type":"string","name":"organisation_name"},{"type":"string","name":"organisation_id"},{"type":"string","name":"private_program"},{"type":"string","name":"project_created_type"},{"type":"string","name":"parent_channel"},{"type":"long","name":"unique_projects"},{"type":"long","name":"unique_users"},{"type":"long","name":"no_of_imp_with_evidence"},{"type":"string","name":"time_stamp"}]},"metricsSpec":[]}}}'
+ml_analytics_druid_distinctCnt_obs_injestion_spec: '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type":"azure","uris":["azure://telemetry-data-store/observation/distinctCount/ml_observation_distinctCount_status.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"ml-obs-status","granularitySpec":{"type":"uniform","queryGranularity":"none","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"time_stamp","format":"auto"},"dimensionsSpec":{"dimensions":[{"type":"string","name":"program_name"},{"type":"string","name":"program_id"},{"type":"string","name":"solution_name"},{"type":"string","name":"solution_id"},{"type":"string","name":"status"},{"type":"string","name":"state_name"},{"type":"string","name":"state_externalId"},{"type":"string","name":"district_name"},{"type":"string","name":"district_externalId"},{"type":"string","name":"organisation_name"},{"type":"string","name":"organisation_id"},{"type":"string","name":"parent_channel"},{"type":"string","name":"solution_type"},{"type":"string","name":"private_program"},{"type":"long","name":"unique_submissions"},{"type":"long","name":"unique_entities"},{"type":"long","name":"unique_users"},{"type":"string","name":"time_stamp"}]},"metricsSpec":[]}}}'
+ml_analytics_druid_distinctCnt_obs_domain_injestion_spec: '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type":"azure","uris":["azure://telemetry-data-store/observation/distinctCount_domain/ml_observation_distinctCount_domain.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"ml-obs-domain","granularitySpec":{"type":"uniform","queryGranularity":"none","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"time_stamp","format":"auto"},"dimensionsSpec":{"dimensions":[{"type":"string","name":"program_name"},{"type":"string","name":"program_id"},{"type":"string","name":"solution_name"},{"type":"string","name":"solution_id"},{"type":"string","name":"state_name"},{"type":"string","name":"state_externalId"},{"type":"string","name":"district_name"},{"type":"string","name":"district_externalId"},{"type":"string","name":"organisation_name"},{"type":"string","name":"organisation_id"},{"type":"string","name":"parent_channel"},{"type":"string","name":"solution_type"},{"type":"string","name":"private_program"},{"type":"string","name":"domain_name"},{"type":"string","name":"domain_externalId"},{"type":"string","name":"domain_level"},{"type":"long","name":"unique_submissions"},{"type":"long","name":"unique_entities"},{"type":"long","name":"unique_users"},{"type":"string","name":"time_stamp"}]},"metricsSpec":[]}}}'
+ml_analytics_druid_distinctCnt_obs_domain_criteria_injestion_spec: '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type":"azure","uris":["azure://telemetry-data-store/observation/distinctCount_domain_criteria/ml_observation_distinctCount_domain_criteria.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"ml-obs-domain-criteria","granularitySpec":{"type":"uniform","queryGranularity":"none","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"time_stamp","format":"auto"},"dimensionsSpec":{"dimensions":[{"type":"string","name":"program_name"},{"type":"string","name":"program_id"},{"type":"string","name":"solution_name"},{"type":"string","name":"solution_id"},{"type":"string","name":"state_name"},{"type":"string","name":"state_externalId"},{"type":"string","name":"district_name"},{"type":"string","name":"district_externalId"},{"type":"string","name":"organisation_name"},{"type":"string","name":"organisation_id"},{"type":"string","name":"parent_channel"},{"type":"string","name":"solution_type"},{"type":"string","name":"private_program"},{"type":"string","name":"domain_name"},{"type":"string","name":"domain_externalId"},{"type":"string","name":"domain_level"},{"type":"string","name":"criteria_name"},{"type":"string","name":"criteria_score"},{"type":"string","name":"criteria_id"},{"type":"long","name":"unique_submissions"},{"type":"long","name":"unique_entities"},{"type":"long","name":"unique_users"},{"type":"string","name":"time_stamp"}]},"metricsSpec":[]}}}'
+ml_analytics_druid_distinctCnt_projects_status_injestion_spec: '{"type":"index","spec":{"ioConfig":{"type":"index","inputSource":{"type":"azure","uris":["azure://telemetry-data-store/projects/distinctCount/ml_projects_distinctCount.json"]},"inputFormat":{"type":"json"}},"tuningConfig":{"type":"index","partitionsSpec":{"type":"dynamic"}},"dataSchema":{"dataSource":"ml-project-status","granularitySpec":{"type":"uniform","queryGranularity":"none","rollup":false,"segmentGranularity":"DAY"},"timestampSpec":{"column":"time_stamp","format":"auto"},"dimensionsSpec":{"dimensions":[{"type":"string","name":"program_name"},{"type":"string","name":"program_id"},{"type":"string","name":"project_title"},{"type":"string","name":"solution_id"},{"type":"string","name":"status_of_project"},{"type":"string","name":"state_name"},{"type":"string","name":"state_externalId"},{"type":"string","name":"district_name"},{"type":"string","name":"district_externalId"},{"type":"string","name":"organisation_name"},{"type":"string","name":"organisation_id"},{"type":"string","name":"private_program"},{"type":"string","name":"project_created_type"},{"type":"string","name":"parent_channel"},{"type":"long","name":"unique_projects"},{"type":"long","name":"unique_users"},{"type":"long","name":"no_of_imp_with_evidence"},{"type":"string","name":"time_stamp"}]},"metricsSpec":[]}}}'
 ml_analytics_obs_distinctCnt_azure_blob_path: "observation/distinctCount/"
 ml_analytics_obs_distinctCnt_domain_azure_blob_path: "observation/distinctCount_domain/"
 ml_analytics_obs_distinctCnt_domain_criteria_azure_blob_path: "observation/distinctCount_domain_criteria/"

ansible/roles/ml-analytics-service/tasks/main.yml

@@ -40,10 +40,23 @@
     - "{{ BASEPATH }}"
     - "{{ WORKDIR }}"
     - "{{ WORKDIR }}/faust_as_service"
-
-- name: Change user and create working directory under opt dir and install python virtual environment
+    
+- name: Delete the virtualenv DIR
+  shell: "rm -rf {{ WORKDIR }}/spark_env"
+  become: true
+  
+- name: Install python virtual environment
   shell: "cd {{ WORKDIR }} && virtualenv --python=python3.8 spark_venv"
   become: true
+  
+- name: Change the ownership of virtual env
+  become: yes
+  file:
+    path: "{{ WORKDIR }}/spark_env"
+    state: directory
+    owner: "{{ USER }}"
+    group: "{{ USER }}"
+    mode: "0755"
 
 - name: Create necessary logs folders for pipeline
   become: yes
@@ -113,4 +126,4 @@
     user: "{{ USER }}"
     minute: "30"
     hour: "18"
-    job: "{{ BASEPATH }}/ml-analytics-service/run.sh"
+    job: "{{ BASEPATH }}/ml-analytics-service/run.sh > {{ BASEPATH }}/ml-analytics-service/crontab_job.log"

ansible/roles/ml-analytics-service/templates/config.j2

@@ -148,14 +148,6 @@ observation_distinctCount_domain_criteria_blob_path = {{ ml_analytics_obs_distin
 
 projects_distinctCnt_blob_path = {{ ml_analytics_projects_distinctCnt_azure_blob_path }}
 
-[REDIS]
-
-host = {{ ml_analytics_redis_host }}
-
-port = {{ ml_analytics_redis_port }}
-
-db_name = {{ ml_analytics_redis_db_name }}
-
 [OUTPUT_DIR]
 
 project = {{ ml_analytics_project_output_dir }}