Saga Server is a JavaSpark server, implementing Saga Library and other tools, as an out-of-the-box question answer system. fully integrated with Saga Admin UI.

Starting the Server

Starting the server is quite easy and can be done with this 2 options:

With built-in configuration (Saga UI compatible)

This execution means using the embedded configuration inside the server which expects a MongoDB called "sagaDB" with the collections and the formats expected from Saga Admin UI.

java -jar -Dfile.encoding=UTF-8 saga-server-parser-0.0.1-SNAPSHOT.jar

With custom configuration and pipeline

This execution requires for a config.js and a pipelines.json file to be passed as parameters of the application, we recommend to overwrite the files within the jar of the server, in order to use and preserve the Saga Admin UI format of the database.

java -jar -Dfile.encoding=UTF-8 saga-server-parser-0.0.1-SNAPSHOT.jar config.json pipelines.json

Server Configuration

Config.js

Bellow is a partial example of the Server Parser config, we omitted the aggregations and fetchTimestamp aggregation from the providers to reduce space in this example:

config.json

{
  "config": {
    "apiPort": 8080,
    "libraryJars": ["./lib"],
    "actionManagerConfig": {
      "actions": "actions-provider:actiongroups"
    },
    "actionProviders": [
      {
        "name": "constant",
        "display": "Constant",
        "type": "Constant",
        "fetchesData": false
      },
      {
        "name": "Human",
        "display": "Human",
        "type": "Elasticsearch",
        "fetchesData": true,
        "transformation": "./transformations/es-transformation.js",
        "hosts": [
          {
            "host": "localhost",
            "port": 9200,
            "schema": "http"
          }
        ],
        "index": "wikidata",
        "query": {
          "bool": {
            "should": [
              {
                "term": {
                  "_id": {
                    "value": "{{human._id}}",
                    "boost": 9999
                  }
                }
              },
              {
                "match_phrase": {
                  "label": {
                    "query": "{{human}}",
                    "boost": 1
                  }
                }
              },
              {
                "match_phrase": {
                  "aliases": {
                    "query": "{{human}}",
                    "boost": 2
                  }
                }
              },
              {
                "query_string": {
                  "query": "{{human.match}}"
                }
              }
            ]
          }
        }
      },
      {
        "name": "geography",
        "display": "Geography",
        "type": "Elasticsearch",
        "fetchesData": true,
        "transformation": "./transformations/es-transformation.js",
        "hosts": [
          {
            "host": "localhost",
            "port": 9200,
            "schema": "http"
          }
        ],
        "index": "wikidata",
        "query": {
          "bool": {
            "should": [
              {
                "term": {
                  "_id": {
                    "value": "{{geography._id}}",
                    "boost": 9999
                  }
                }
              },
              {
                "match_phrase": {
                  "label": {
                    "query": "{{geography}}",
                    "boost": 1
                  }
                }
              },
              {
                "match_phrase": {
                  "aliases": {
                    "query": "{{geography}}",
                    "boost": 2
                  }
                }
              },
              {
                "query_string": {
                  "query": "{{geography.match}}"
                }
              }
            ]
          }
        }
      },
      {
        "name": "currency",
        "display": "Currency",
        "type": "Openexchange",
        "transformation": "./transformations/openexchange-transformation.js",
        "fetchesData": false,
        "appId": "6cfdc0df634a4796b1a40748cdbe6006"
      }
    ],
    "providers": [
      {
        "name": "actions-provider",
        "type": "MongoDB",
        "uri": "mongodb://localhost:27017",
        "database": "sagaDB",
        "aggregation": [...],
        "transactionCollection": "transactions",
        "fetchTimestamp": [...]
      },
      {
        "name": "entity-provider",
        "type": "MongoDB",
        "uri": "mongodb://localhost:27017",
        "database": "sagaDB",
        "aggregation": [...],
        "transactionCollection": "transactions",
        "fetchTimestamp": [...]
      },
      {
        "name": "patterns-provider",
        "type": "MongoDB",
        "uri": "mongodb://localhost:27017",
        "database": "sagaDB",
        "aggregation": [...],
        "transactionCollection": "transactions",
        "fetchTimestamp": [...]
      },
      {
        "name": "regex-provider",
        "type": "MongoDB",
        "uri": "mongodb://localhost:27017",
        "database": "sagaDB",
        "aggregation": [...],
        "transactionCollection": "transactions",
        "fetchTimestamp": [...]
      },
	  {
        "name": "equipment-provider",
        "type": "MongoDB",
        "uri": "mongodb://localhost:27017",
        "database": "sagaDB",
        "aggregation": [...],
        "transactionCollection": "transactions",
        "fetchTimestamp": [...]
      },
	  {
        "name": "unit-provider",
        "type": "MongoDB",
        "uri": "mongodb://localhost:27017",
        "database": "sagaDB",
        "aggregation": [...],
        "transactionCollection": "transactions",
        "fetchTimestamp": [...]
      }
    ]
  }
}

apiPort - port in which the server will listen for request
libraryJars - paths to the folders holding the jar which need to be added to the classpath. The paths can be either absolute or relative
actionManagerConfig - Configuration of the Action Manager, it decides which action providers needs to handle the graph response
- actions - call the resource provider holding the action definitions.
actionProviders - A list of the implemented action providers to be use in the Action Manager. Each action definition can be found in Saga Actions
providers - Standar resource providers, the provider holding the actions is declared in this section also.

Pipelines.js

The other file part of the configuration is the pipelines.json files, which holds the pipelines definition, currently we have 2 pipelines for Saga Parser:

process - is defined for text processing
ml - used as a machine learning preprocess for the Entity Trainer

Besides that, the pipelines are exactly as decribe in Configure Pipelines & Resource Providers

pipelines.json

{
  "config": {
    "pipelineConfiguration": {
      "process": {
        "reader": {
          "type": "SimpleReader",
          "splitRegex": "[\r\n]+"
        },
        "stages": [
          {
            "type": "SentenceBreakerStage"
          },
          {
            "type": "WhitespaceTokenizerStage",
            "requiredFlags": [
              "SENTENCE"
            ]
          },
          {
            "type": "CharacterSplitter"
          },
          {
            "type": "CharChangeSplitter",
            "case": true,
            "number": true,
            "punctuation": true
          },
          {
            "type": "LemmatizeStage",
            "exclude": [
              "ob",
              "syn",
              "alt"
            ],
            "skipFlags": [
              "ALL_PUNCTUATION"
            ]
          },
          {
            "type": "CaseAnalysisStage"
          },
          {
            "type": "NumberRecognizer"
          },
          {
            "type": "StopWordsStage"
          },
          {
            "type": "RegexPatternStage",
            "patterns": "regex-provider:patterns",
            "caseInsensitive": true,
            "boundaryFlags": [
              "SENTENCE_SPLIT"
            ]
          },
          {
            "type": "DictionaryTaggerStage",
            "dictionary": "entity-provider:entities",
            "skipFlags": [
              "SKIP"
            ],
            "boundaryFlags": [
              "SENTENCE_SPLIT"
            ],
            "requiredFlags": [
              "TOKEN",
              "ALL_LOWER_CASE"
            ],
            "ignoreTags": [
              "root"
            ],
            "debug": true
          },
          {
            "type": "AdvancedPattern",
            "skipFlags": [
              "SKIP"
            ],
            "patterns": "patterns-provider:patterns",
            "debug": true
          }
        ]
      },
      "ml": {
        "reader": {
          "type": "SimpleReader",
          "splitRegex": "[\r\n]+"
        },
        "stages": [
          {
            "type": "QuotationBreakerStage",
            "singleQuotes": true
          },
          {
            "type": "SentenceBreakerStage",
            "skipFlags": [
              "PROCESSED"
            ]
          },
          {
            "type": "WhitespaceTokenizerStage",
            "requiredFlags": [
              "SENTENCE"
            ]
          },
          {
            "type": "CharChangeSplitter",
            "case": false,
            "number": true,
            "punctuation": true
          },
          {
            "type": "CaseAnalysisStage"
          },
          {
            "type": "StopWordsStage"
          },
          {
            "type": "RegexPatternStage",
            "patterns": "regex-provider:patterns",
            "caseInsensitive": true,
            "boundaryFlags": [
              "SENTENCE_SPLIT"
            ]
          },
          {
            "type": "DictionaryTaggerStage",
            "dictionary": "unit-provider:entities",
            "skipFlags": [
              "SKIP"
            ],
            "boundaryFlags": [
              "SENTENCE_SPLIT"
            ],
            "requiredFlags": [
              "TOKEN",
              "ALL_LOWER_CASE"
            ],
            "ignoreTags": [
              "root"
            ]
          },
          {
            "type": "AdvancedPattern",
            "skipFlags": [
              "SKIP"
            ],
            "patterns": "patterns-provider:patterns"
          },
          {
            "type": "DictionaryTaggerStage",
            "dictionary": "equipment-provider:entities",
            "skipFlags": [
              "SKIP"
            ],
            "boundaryFlags": [
              "SENTENCE_SPLIT"
            ],
            "requiredFlags": [
              "TOKEN",
              "ALL_LOWER_CASE"
            ],
            "ignoreTags": [
              "root"
            ]
          }
        ]
      }
    }
  }
}

The default pipelines configuration can change in time, with the creation of new stages and the improvements to the code

Page tree

Saga Server