Compare schemas

This is the official documentation of the @forestadmin/agent Node.js agent.

Checklist

Checking that the agents are compatible

Making an agent which works is not enough: you also need to make sure that it generates a schema where the naming of most entities is the same as the old one.

This is because all the UI / and security customizations that you have made are identified by the name of the entity to which they are applied.

This is valid for:

• Charts (in dashboards and the collection view)

• Order and visibility of fields in table-view, detail-form, related data, segments

• Roles and Scopes

• Segments visibility

• Summary views and Workspaces

Example

Broken chart after migration

For instance, if you created a chart on the customers collection that displays the TOP 10 customers by number of orders, you will need to make sure that:

• The customers collection still exists and has the same name

• The orders collection still exists and has the same name

• The customers collection still has a relationship called orders

Checking the compatibility with .forestadmin-schema.json, jq and diff

• jq is a command line tool that allows you to manipulate JSON files. If you don't have it installed, you can install it with one of the following methods ↗ or use it online ↗.
• diff is a command line tool that allows you to compare two files. It is installed by default on virtually all Unix systems, windows users can use online versions ↗.

To ensure agents are compatible, we have a tool at our disposal: the .forestadmin-schema.json file.

However, diffing two .forestadmin-schema.json files is extremely tedious: the file contains a lot of information that is not relevant to our objective.

Instead, we can choose to focus on the differences that matter the most: the name of the different entities.

Extracting the list of entities

Extracting the list of entities exported by an agent can be done with the following jq command.

Run this command in the directory of both your old and new agent: it will generate a file called agent-entities.txt in the current directory.

$ jq -r '.collections[] | (
    "Collection: \"\(.name)\"",
    "Action: \"\(.name)\" has \"\(.actions[].name)\"",
    "Field: \"\(.name)\" has \"\(.fields[] | select(.reference == null) | .field)\"",
    "Relationship: \"\(.name)\" has \"\(.fields[] | select(.reference) | .field)\"",
    "Segment: \"\(.name)\" has \"\(.segments[].name)\""
)' .forestadmin-schema.json | sort > agent-entities.txt

Equivalent Node.js script

You can also execute the following script to generate the same file:

const fs = require('fs');
const forestAdminSchema = require('./.forestadmin-schema.json');

const entities = forestAdminSchema.collections
  .map(collection => [
    `Collection: "${collection.name}"`,
    ...collection.actions.map(
      action => `Action: "${collection.name}" has "${action.name}"`,
    ),
    ...collection.fields
      .filter(field => !field.reference)
      .map(field => `Field: "${collection.name}" has "${field.field}"`),
    ...collection.fields
      .filter(field => field.reference)
      .map(field => `Relationship: "${collection.name}" has "${field.field}"`),
    ...collection.segments.map(
      segment => `Segment: "${collection.name}" has "${segment.name}"`,
    ),
  ])
  .flat()
  .sort();

fs.writeFileSync('agent-entities.txt', entities.join('\n'));

Comparing the list of entities

Once you have generated the agent-entities.txt file for both your old and new agent, you can diff them with the following command:

$ diff --side-by-side --suppress-common-lines --minimal [old-agent-folder]/agent-entities.txt [new-agent-folder]/agent-entities.txt
Collection: "account"                    <
Field: "account" has "_id"               <
Field: "account" has "firstname"         <
Field: "account" has "lastname"          <
Field: "account_address" has "_id"       <
Field: "account_address" has "country"   | Field: "account_address" has "Country"
Relationship: "account" has "address"    <
Relationship: "account" has "bills"      <
Relationship: "account" has "store"      <

Diff files are read using the following format:

• >: line is present in the new file, but not in the old one

• <: line is present in the old file, but not in the new one

• '|': line is present in both files, but the content is different

Fixing differences

Once you have identified the differences, you can fix them by:

Danger level	Difference	Fix
🔴	Collections or fields have different names (i.e. camelCase vs snake_case)	Rename the collections in the new agent (docs)
🔴	Missing Smart features (actions, fields, relationships, segments, ...)	Port the missing smart features to your new agent
🟢	Extra collections	You can remove them (docs)
🟢	Extra fields	You can remove them (docs)