Here I wanted more consistency between different records.
This is the typical layout for any record. Not all records have all these fields.
{
_id: <ObjectId>, // Always auto-set at creation time.
recordType: <string>, // e.g. component, form, test, job
collection: <string>, // collection name
componentUuid: <BSON object>, // UUID of component, binary form.
formId: <string>, // id of form used
data: {} // for tests and jobs and components
metadata: {} // reserved for formio stuff
schema: {} // for forms
}
All records (forms, components, test, jobs) should have this block.
{
insertDate: <date>, // Timestamp written to database
ip: <string> // ip address of creator client
user: <user> // trimmed user record (see below)
}
Used in the Insertion block, above.
user: {
displayName: <String>, // Printable name
user_id: <string>, // auth0 user_id
emails: [ <string>, ...] // email list
}
All records that evolve (components, forms) should have this.
validity: {
startDate: <Date> // Date this version becomes active
version: <integer> // version number for validity; later numbers are more correct
ancestor_id: <ObjectId> // _id of the version edited to create this one.
}
{
createdFrom: // Exists only if this record was created by process
processRecordId: <ObjectId> // Record ID 'processed' collection
id: <ObjectID>, // The record (typically a job) that led to the creation of this object
recordType: <String>, // usually a job
collection: <String>, // usually 'jobs'
formId: <String> // Name of the form that did this
formObjectId: <ObjectId> // specific reference to the form that did this.
processId: <String>, // Name of the process invoked.
}
}
All records in the datbase should have the following fields:
_id: <ObjectId>, // Auto-set by Mongo, Includes insertion timestamp redundantly.
recordType: <string>, // component, form, test, job
(Because the id field is set by Mongo, it contains a timestamp. This timestamp should be the same as the insertion.startDate timestamp to within a second.)
In addition, some will have this field:
collection: <string> // the name of the Mongo collection this record is stored in
{
// supplied by API:
_id: <ObjectId>, // Auto-set by Mongo, Includes insertion timestamp redundantly.
recordType: "component", // component, form, test, job
insertion: <insertion block>,
state: "submitted",
referencesComponents: [ <BSON Uuid>,... ], // list of uuids in data given below.., found by deep searh
createdFrom: {}, //optional, see createdFrom block.
// supplied by caller:
type: <string>,
componentUuid, // component uuid
validity: <validity block> // Fields are auto-set if empty.
data: {
name: <string> // user-readable name
}
metadata: {}
}
{
// supplied by API:
_id: <ObjectId>, // Auto-set by Mongo, Includes insertion timestamp redundantly.
recordType: "form", // component, form, test, job
collection: <string>, // which form collection this comes from.
insertion: <insertion block>
// Supplied by caller:
validity: <validity block>
formId: <string> // identifier for this form type
formName: <string> // Name of this form
icon: <file reference>, // nice icon to display this thing with.
tags: <array of string> // Category tags. "Trash" = don't show this entry to users
componentType: <string> // Applies to Test forms only; limits which Components have this as a test option.
schema: {
components:[...]
}
processes: { // optional
<processId> : <algorithm,
...
}
}
{
// supplied by API:
_id: <ObjectId>, // Auto-set by Mongo, Includes insertion timestamp redundantly.
recordType: <string> // "test" or "job" respectively
insertion: <insertion block>
createdFrom: {}, //optional, see createdFrom block.
}]
// supplied by caller:
componentUuid, // BSON component UUID. Required for 'test', should not be there for 'job'. Provided via api route.
formId: <string>,
formName: <string>, // OMIT???
icon: <file reference>, // nice icon to display this thing with.
tags: <array of string> // Category tags. "Trash" = don't show this entry to users
formObjectId: <ObjectId>, // objectID of the form record used.
state: <string> // Required. "submitted" for final data, "draft" for a draft
// Also reserved: 'trash'
data: { ... } // actual test data. (Also contains uuid?)
metadata: { ... }, // optional, Formio stuff.
processed: [<objects>] // processes that have been applied to record
}
Collection 'processed'.
{
_id: <ObjectId>
input: { // Describes the original Job record that was procssed/
recordType: <String>,
collection: <String>,
_id: <ObjectId>
},
process: {
_id: <ObjectId>
formId: <String>
collection: <String>,
processId: <String>
}
}
This collection is used to find connections between Components, and is reconstructed automatically by either map-reduce or on record insert. This is not structured like the others, since it is basically a reconstructable index rather than storage.
{
_id: <BSON Uuid>,
referredToBy: [<BSON Uuid>,... ]
}
This represents a sequence of tasks that should be taken in terms of work proceedures. For example, an element might need to be entered in the DB, then located, then weighed, then tested for tension, then shipped.
{
_id: <BSON Uuid>, // specific object
recordType: "course",
courseId: <string>, // unique name of this path
validity: <validity>, // This is a versioned object
insertion: <insertion>,
name: <string>, // user-readable name of this path (mutable)
icon: <file reference>, // nice icon to display this thing with.
tags: [ <string>,... ], // Which tags to apply. Typically only one tag
path: [ <step>, <step>, <step>],
componentType, // Primary component type this course will track.
};
where
<path> = {
[<step>,...]
}
and
<step> = {
type: 'component|job|test',
formId: <string>,
advice: <string> , // Mouse-over or other explanitory text
identifier: <string>, // valid for component and job types.
// Dot-notation version of how to look up the object in question in the record. i.e. if
// Workflow has an apaId value that stores the UUID, this should be "data.uuid". Does
// not yet support array values.
}
Wiki or Doc entries are tracked like Components: there is a single ID (alphanumeric for the url) which identifies each, and old versions are kept.
{
_id: <BSON Uuid>, // specific object
recordType: "doc",
docId: <string>, // unique name of this path
validity: <validity>, // This is a versioned object
insertion: <insertion>, // Last insertion
data: <markdown string>,
};