coco.yml

env:
  ES_ENDPOINT: https://localhost:9200
  ES_USERNAME: admin
  ES_PASSWORD: $[[keystore.ES_PASSWORD]]
  WEB_BINDING: 0.0.0.0:9000
  API_BINDING: 0.0.0.0:2900

coco:
  server:
    public: false
    name: "My Coco Server"
    encode_icon_to_base64: false # enable this if coco-server is using self-signed certs
#    #[START] for managed mode only, need to configured by coco-server operator
#    endpoint: "http://localhost:9001/"
#    auth_provider: #config for app to use
#      sso:
#        url: "http://localhost:9001/#/login"
#    #[END] for managed mode only
    minimal_client_version:
      number: "0.3"
    provider:
      name: "INFINI Labs"
      description: "Coco AI Server - Search, Connect, Collaborate, AI-powered enterprise search, all in one space."
      icon: "https://coco.rs/favicon.ico"
      website: "https://coco.rs/"
      eula: "https://coco.rs/#/terms"
      privacy_policy: "https://coco.rs/privacy"
      banner: "https://coco.rs/svg/connect.svg"
      #[START] for managed mode only, config for coco-server to use
      auth_provider:
        sso:
          url: "/sso/login/cloud?provider=coco-cloud&product=coco"
      #[END] for managed mode only

## Cleanup exists data:
# curl -u admin:45ff432a5428ade77c7b -k -XDELETE   https://localhost:9200/_template/coco-search
# curl -u admin:45ff432a5428ade77c7b -k -XDELETE   https://localhost:9200/_template/coco
# curl -u admin:45ff432a5428ade77c7b -k -XDELETE  https://localhost:9200/_scripts/coco-query-string
# curl -u admin:45ff432a5428ade77c7b -k -XDELETE  https://localhost:9200/_scripts/coco-query-string-extra-should
# curl -u admin:45ff432a5428ade77c7b -k -XDELETE   https://localhost:9200/coco_*
elastic:
  elasticsearch: prod
  enabled: true
  store:
    enabled: false
  orm:
    enabled: true
    index_prefix: "coco_"
    override_exists_template: true
    index_templates:
      "coco-search": |
        {
          "order": 10,
          "index_patterns": [
            "coco_*"
          ],
          "settings": {
            "index": {
              "analysis": {
                "char_filter": {
                    "tsconvert" : {
                        "type" : "stconvert",
                        "delimiter" : ",",
                        "keep_both" : false,
                        "convert_type" : "t2s"
                    }
                 },
                "tokenizer": {
                      "my_pinyin": {
                        "type": "pinyin",
                        "keep_first_letter": true,
                        "keep_separate_first_letter": true,
                        "keep_full_pinyin": true,
                        "keep_original": false,
                        "limit_first_letter_length": 16,
                        "lowercase": true
                      }
                    },
                "analyzer": {
                  "pinyin_analyzer": {
                     "tokenizer": "my_pinyin"
                   },
                  "combined_text_analyzer": {
                    "char_filter": ["tsconvert"],
                    "filter": [
                      "lowercase",
                      "asciifolding",
                      "word_delimiter",
                      "unique"
                    ],
                    "tokenizer": "classic"
                  }
                }
              }
            }
          },
          "mappings": {}
        }

    search_templates:
      "coco-query-string": |
        {
            "from": "{{from}}",
            "size": "{{size}}",
            "_source": {
                "includes": [ "{{#source}}","{{.}}","{{/source}}"],
                "excludes": [ "payload.*","content" ]
            },
            "query": {
              "bool": {
                "must": {{#toJson}}must_clauses{{/toJson}},
                "must_not": [],
                "should": [
                  {{#query}}
                  {
                    "prefix": {
                      "{{field}}.keyword": {
                        "value": "{{query}}",
                        "boost": 100
                      }
                    }
                  },
                  {
                    "match_phrase_prefix": {
                      "{{field}}": {
                        "query": "{{query}}",
                        "boost": 50
                      }
                    }
                  },
                  {
                    "match_phrase": {
                      "{{field}}": {
                        "query": "{{query}}",
                        "boost": 30
                      }
                    }
                  },
                  {
                    "match": {
                      "{{field}}": {
                        "query": "{{query}}",
                        "fuzziness": "AUTO",
                        "max_expansions": 10,
                        "boost": 5,
                        "prefix_length": 2,
                        "operator": "and",
                        "fuzzy_transpositions": true
                      }
                    }
                  },
                  {
                    "query_string": {
                      "fields": ["{{field}}","{{field}}.keyword^10","combined_fulltext"],
                      "query": "{{query}}",
                      "fuzziness": "AUTO",
                      "fuzzy_prefix_length": 2,
                      "fuzzy_max_expansions": 10,
                      "boost": 3,
                      "default_operator": "AND",
                      "fuzzy_transpositions": true,
                      "allow_leading_wildcard": false
                    }
                  },{"match": {
                      "{{field}}.pinyin": {
                          "query": "{{query}}",
                          "boost": 4
                          }
                    }}
                  {{/query}}
                ]
              }
            }
        }
      "coco-query-string-extra-should": |
        {
            "from": "{{from}}",
            "size": "{{size}}",
            "_source": {
                "includes": [ "{{#source}}","{{.}}","{{/source}}"],
                "excludes": [ "payload.*","content" ]
            },
            "query": {
              "bool": {
                "must": {{#toJson}}must_clauses{{/toJson}},
                "must_not": [],
                "should": [
                  {{#query}}
                  {
                    "prefix": {
                      "{{field}}.keyword": {
                        "value": "{{query}}",
                        "boost": 100
                      }
                    }
                  },
                  {
                    "match_phrase_prefix": {
                      "{{field}}": {
                        "query": "{{query}}",
                        "boost": 50
                      }
                    }
                  },
                  {
                    "match_phrase": {
                      "{{field}}": {
                        "query": "{{query}}",
                        "boost": 30
                      }
                    }
                  },
                  {
                    "match": {
                      "{{field}}": {
                        "query": "{{query}}",
                        "fuzziness": "AUTO",
                        "max_expansions": 10,
                        "prefix_length": 2,
                        "boost": 5,
                        "fuzzy_transpositions": true
                      }
                    }
                  },
                  {
                    "query_string": {
                      "fields": ["{{field}}","{{field}}.keyword^10","combined_fulltext"],
                      "query": "{{query}}",
                      "fuzziness": "AUTO",
                      "fuzzy_prefix_length": 2,
                      "default_operator": "AND",
                      "boost": 3,
                      "fuzzy_max_expansions": 10,
                      "fuzzy_transpositions": true,
                      "allow_leading_wildcard": false
                    }
                  },{"match": {
                    "{{field}}.pinyin": {
                        "query": "{{query}}",
                        "boost": 4
                        }
                    }}
                  {{/query}}
                   ,{{#toJson}}extra_should_clauses{{/toJson}}
                ]
              }
            }
        }

elasticsearch:
  - name: prod
    enabled: true
    endpoints:
      - $[[env.ES_ENDPOINT]]
    discovery:
      enabled: false
    basic_auth:
      username: $[[env.ES_USERNAME]]
      password: $[[env.ES_PASSWORD]]

api:
  enabled: true
  websocket:
    enabled: false
  network:
    binding: $[[env.API_BINDING]]

web:
  enabled: true
  embedding_api: false
  network:
    binding: $[[env.WEB_BINDING]]
  tls:
    enabled: false
    skip_insecure_verify: true
    default_domain: "localhost:2900"
    #    default_domain: "api.coco.rs"
    auto_issue:
      enabled: false
      email: "hello@infinilabs.com"
      include_default_domain: true
      domains:
        - "www.coco.rs"
      provider:
        tencent_dns:
          secret_id: $[[keystore.TENCENT_DNS_ID]] #./bin/coco keystore add TENCENT_DNS_ID
          secret_key: $[[keystore.TENCENT_DNS_KEY]] #./bin/coco keystore add TENCENT_DNS_KEY
  security:
    enabled: true
    managed: false #remote managed instance
    authentication:
      oauth:
        cloud:
          enabled: true
          client_secret: "999999"
          authorize_url: "http://localhost:9000/oauth/authorize"
          token_url: "http://localhost:9000/oauth/token"
          profile_url: "http://localhost:9000/account/profile"
          redirect_url: "http://localhost:9001/sso/callback/cloud"
          success_page: "/#/home"
#          failed_page: "/#/user/sso/failed"
          scopes: [ "openid","email","profile"]


connector:
  google_drive: # get your token here: https://developers.google.com/drive/api/quickstart/go
    enabled: true
    queue:
      name: indexing_documents
    interval: 10s
    skip_invalid_token: true
  yuque:
    enabled: true
    queue:
      name: indexing_documents
    interval: 10s
  hugo_site:
    enabled: true
    interval: 10s
    queue:
      name: indexing_documents
  notion:
    enabled: true
    interval: 10s
    queue:
      name: indexing_documents
  rss:
    enabled: true
    interval: 10s
    queue:
      name: indexing_documents
  local_fs:
    enabled: true
    interval: 10s
    queue:
      name: indexing_documents
  s3:
    enabled: true
    interval: 30s
    queue:
      name: indexing_documents
  confluence:
    enabled: true
    interval: 5m
    queue:
      name: indexing_documents
  network_drive:
    enabled: true
    interval: 30s
    queue:
      name: indexing_documents

##background jobs
pipeline:
  - name: enrich_documents
    auto_start: false
    keep_running: true
    processor:
      - consumer:
          auto_commit_offset: true
          queue_selector:
            keys:
              - indexing_documents
          consumer:
            group: enriched_documents
            fetch_max_messages: 10
          processor:
            - document_enrichment:
                model: $[[env.ENRICHMENT_MODEL]]
                input_queue: "indexing_documents"
                min_input_document_length: 500
                output_queue:
                  name: "enriched_documents"
                  label:
                    tag: "enriched"

  - name: merge_documents
    auto_start: true
    keep_running: true
    processor:
      - indexing_merge:
          input_queue: "indexing_documents"
#          input_queue: "enriched_documents"
          idle_timeout_in_seconds: 1
          elasticsearch: "prod"
          index_name: "coco_document-v2"
          key_field: "id"
          output_queue:
            name: "merged_documents"
            label:
              tag: "merged"
          worker_size: 1
          bulk_size_in_kb: 10240
  - name: ingest_documents
    auto_start: true
    keep_running: true
    processor:
      - bulk_indexing:
          bulk:
            compress: true
            batch_size_in_mb: 10
            batch_size_in_docs: 10240
          consumer:
            fetch_max_messages: 100
          queues:
            type: indexing_merge
            tag: "merged"


http_client:
  default:
    proxy:
      enabled: false
      default_config:
        http_proxy: http://127.0.0.1:7890
        socket5_proxy: socks5://127.0.0.1:7890
      override_system_proxy_env: true #override system proxy's environment settings with this proxy settings
      permitted:
        - "google.com"
      denied:
        - "localhost"
        - "www.yuque.com"
        - "yuque.com"
        - "infinilabs.com"
        - "api.coco.rs"
        - "dashscope.aliyuncs.com"
      domains:
        "github.com":
          http_proxy: http://127.0.0.1:7890
          socket5_proxy: socks5://127.0.0.1:7890