Prototyping new json schema

[DSuveges]

As part of the consolidation of the evidence objects in the backend, we are re-modeling the json schema to reflect the new simplified/flattened design.

Based on the meeting we had on 2020.11.11 the following conclusions were reached:

• We need to maintain a json schema that guides our data providers and can be used as template to generate evidence strings.
• The schema will reflect the concepts of the new platform design, so the units of the schema is going to be data source centric instead of data type.
• Each of the valuable columns will be defined in a common section.
• For each data source there will only be a list of required fields.
• We haven't reached a consensus on how the unique association fields are defined, and at which point of the evidence generation. So for the first iteration of the json schema, the unique_association_fields will be omitted.

The schema is written based on the most recent iteration of the evidence schema review document.

[ireneisdoomed]

As discussed on Nov 12th daily meeting we should keep an eye on:

• Fields present in the submitted evidence that were not taken into account in the cleaned up version.
• Check if the scoring depends on fields that are missing.
• See which data sources provide PubMed Central IDs
• Maintain a document for those questions that arise and we want to ask the data providers

[DSuveges]

Structured fields:

• literature: list of pmids/pmc ids

|-- literature: array (nullable = true)
| |-- element: string (containsNull = true)

• diseaseModelAssociatedModelPhenotypes: list of objects:

|-- diseaseModelAssociatedModelPhenotypes: array (nullable = true)
| |-- element: struct (containsNull = true)
| | |-- id: string (nullable = true)
| | |-- label: string (nullable = true)

• diseaseModelAssociatedHumanPhenotypes: list of objects:

|-- diseaseModelAssociatedHumanPhenotypes: array (nullable = true)
| |-- element: struct (containsNull = true)
| | |-- id: string (nullable = true)
| | |-- label: string (nullable = true)

• variations: list of objects

|-- variations: array (nullable = true)
| |-- element: struct (containsNull = true)
| | |-- functionalConsequenceId: string (nullable = true)
| | |-- inheritancePattern: string (nullable = true)
| | |-- numberMutatedSamples: long (nullable = true)
| | |-- numberSamplesTested: long (nullable = true)
| | |-- numberSamplesWithMutationType: long (nullable = true)
| | |-- variantAminoacidDescription: string (nullable = true)

• textMiningSentences: list of objects:

|-- textMiningSentences: array (nullable = true)
| |-- element: struct (containsNull = true)
| | |-- dEnd: long (nullable = true)
| | |-- dStart: long (nullable = true)
| | |-- section: string (nullable = true)
| | |-- tEnd: long (nullable = true)
| | |-- tStart: long (nullable = true)
| | |-- text: string (nullable = true)

• signigicantDriverMethod: list of string:

|-- significantDriverMethods: array (nullable = true)
| |-- element: string (containsNull = true)

• clinicalUrls: list of objects:

|-- clinicalUrls: array (nullable = true)
| |-- element: struct (containsNull = true)
| | |-- niceName: string (nullable = true)
| | |-- url: string (nullable = true)

[DSuveges]

As for the first round, I close the issue. The updates are not yet merged to the master branch.
Updates available: https://github.com/opentargets/json_schema/tree/ds_1249_new_json_schema