Merged in feature/splitqueryrunning (pull request #57)

Split Query Running + Debugging Full Flow

* completedquerysyncrunner

* spliitinglogic

* synccomplete

* informdependents

* only push same collector

* deps

* livetesting

* foundissue

* some issues resolved

* activeupdate

* collectorupdatefixes

* fix dbquesries

* tests

* tests

* pollingdebug
This commit is contained in:
Michael McGuinness
2025-02-11 15:22:59 +00:00
parent 24a038ec3d
commit 71f9802e1a
108 changed files with 3013 additions and 1666 deletions
+142 -105
View File
@@ -45,76 +45,122 @@ func (q *Queries) GetResultValueWithVersion(ctx context.Context, arg *GetResultV
}
const listQueryRequirementValues = `-- name: ListQueryRequirementValues :many
WITH latest_versions AS (
SELECT
r.queryId,
MAX(r.cleanVersion) as max_clean_version,
MAX(r.textVersion) as max_text_version
FROM results r
WHERE r.documentId = $3
AND r.queryVersion = $2
GROUP BY r.queryId
)
SELECT rq.requiredQueryId as queryId, r.value, q.type
WITH reqQueries as (
SELECT q.id as queryId, q.activeVersion, q.type
FROM requiredQueries as rq
JOIN results as r on r.queryId = rq.requiredQueryId
JOIN queries as q on q.id = rq.requiredQueryId
JOIN latest_versions lv ON lv.queryId = r.queryId
WHERE rq.queryId = $1 and r.documentId = $3
and $2 >= rq.addedVersion
and $2 < COALESCE(rq.removedVersion, $2 + 1)
and r.queryVersion = $2
and r.cleanVersion >= $4 and r.textVersion >= $5
AND r.cleanVersion = lv.max_clean_version
AND r.textVersion = lv.max_text_version
WHERE rq.queryId = $2
and $3 >= rq.addedVersion
and (rq.removedVersion is null or $3 < rq.removedVersion)
),
codeVersions as (
SELECT
d.id as documentId,
coalesce(ccv.minCleanVersion, 1) as minCleanVersion,
coalesce(ccv.minTextVersion, 1) as minTextVersion
FROM documents as d
LEFT JOIN collectors as c on c.jobId = d.jobId
LEFT JOIN collectorCodeVersions as ccv on c.id = ccv.collectorId
and c.activeVersion >= ccv.addedVersion
and c.activeVersion < COALESCE(ccv.removedVersion, c.activeVersion)
WHERE d.id = $1
LIMIT 1
),
latestVersions AS (
SELECT
rq.queryId,
rq.type,
r.queryVersion,
r.cleanVersion,
r.textVersion,
ROW_NUMBER() OVER (
PARTITION BY rq.queryId
ORDER BY r.cleanVersion DESC, r.textVersion DESC
) as rowNumber
FROM reqQueries as rq
LEFT JOIN results as r ON rq.queryId = r.queryId
and r.documentId = $1
JOIN codeVersions as ccv on ccv.documentId = r.documentId
WHERE r.documentId = $1
and r.queryVersion = rq.activeVersion
and r.cleanVersion >= ccv.minCleanVersion
and r.textVersion >= ccv.minTextVersion
)
SELECT lv.queryId, lv.type, r.value
FROM latestVersions as lv
JOIN results as r ON r.queryId = lv.queryId
and r.documentId = $1
and r.queryVersion = lv.queryVersion
and r.cleanVersion = lv.cleanVersion
and r.textVersion = lv.textVersion
and lv.rowNumber = 1
`
type ListQueryRequirementValuesParams struct {
Queryid pgtype.UUID `db:"queryid"`
Addedversion int32 `db:"addedversion"`
Documentid pgtype.UUID `db:"documentid"`
Cleanversion int32 `db:"cleanversion"`
Textversion int32 `db:"textversion"`
Documentid pgtype.UUID `db:"documentid"`
Queryid pgtype.UUID `db:"queryid"`
Version int32 `db:"version"`
}
type ListQueryRequirementValuesRow struct {
Queryid pgtype.UUID `db:"queryid"`
Value string `db:"value"`
Type Querytype `db:"type"`
Value string `db:"value"`
}
// ListQueryRequirementValues
//
// WITH latest_versions AS (
// SELECT
// r.queryId,
// MAX(r.cleanVersion) as max_clean_version,
// MAX(r.textVersion) as max_text_version
// FROM results r
// WHERE r.documentId = $3
// AND r.queryVersion = $2
// GROUP BY r.queryId
// )
// SELECT rq.requiredQueryId as queryId, r.value, q.type
// WITH reqQueries as (
// SELECT q.id as queryId, q.activeVersion, q.type
// FROM requiredQueries as rq
// JOIN results as r on r.queryId = rq.requiredQueryId
// JOIN queries as q on q.id = rq.requiredQueryId
// JOIN latest_versions lv ON lv.queryId = r.queryId
// WHERE rq.queryId = $1 and r.documentId = $3
// and $2 >= rq.addedVersion
// and $2 < COALESCE(rq.removedVersion, $2 + 1)
// and r.queryVersion = $2
// and r.cleanVersion >= $4 and r.textVersion >= $5
// AND r.cleanVersion = lv.max_clean_version
// AND r.textVersion = lv.max_text_version
// WHERE rq.queryId = $2
// and $3 >= rq.addedVersion
// and (rq.removedVersion is null or $3 < rq.removedVersion)
// ),
// codeVersions as (
// SELECT
// d.id as documentId,
// coalesce(ccv.minCleanVersion, 1) as minCleanVersion,
// coalesce(ccv.minTextVersion, 1) as minTextVersion
// FROM documents as d
// LEFT JOIN collectors as c on c.jobId = d.jobId
// LEFT JOIN collectorCodeVersions as ccv on c.id = ccv.collectorId
// and c.activeVersion >= ccv.addedVersion
// and c.activeVersion < COALESCE(ccv.removedVersion, c.activeVersion)
// WHERE d.id = $1
// LIMIT 1
// ),
// latestVersions AS (
// SELECT
// rq.queryId,
// rq.type,
// r.queryVersion,
// r.cleanVersion,
// r.textVersion,
// ROW_NUMBER() OVER (
// PARTITION BY rq.queryId
// ORDER BY r.cleanVersion DESC, r.textVersion DESC
// ) as rowNumber
// FROM reqQueries as rq
// LEFT JOIN results as r ON rq.queryId = r.queryId
// and r.documentId = $1
// JOIN codeVersions as ccv on ccv.documentId = r.documentId
// WHERE r.documentId = $1
// and r.queryVersion = rq.activeVersion
// and r.cleanVersion >= ccv.minCleanVersion
// and r.textVersion >= ccv.minTextVersion
// )
// SELECT lv.queryId, lv.type, r.value
// FROM latestVersions as lv
// JOIN results as r ON r.queryId = lv.queryId
// and r.documentId = $1
// and r.queryVersion = lv.queryVersion
// and r.cleanVersion = lv.cleanVersion
// and r.textVersion = lv.textVersion
// and lv.rowNumber = 1
func (q *Queries) ListQueryRequirementValues(ctx context.Context, arg *ListQueryRequirementValuesParams) ([]*ListQueryRequirementValuesRow, error) {
rows, err := q.db.Query(ctx, listQueryRequirementValues,
arg.Queryid,
arg.Addedversion,
arg.Documentid,
arg.Cleanversion,
arg.Textversion,
)
rows, err := q.db.Query(ctx, listQueryRequirementValues, arg.Documentid, arg.Queryid, arg.Version)
if err != nil {
return nil, err
}
@@ -122,7 +168,7 @@ func (q *Queries) ListQueryRequirementValues(ctx context.Context, arg *ListQuery
items := []*ListQueryRequirementValuesRow{}
for rows.Next() {
var i ListQueryRequirementValuesRow
if err := rows.Scan(&i.Queryid, &i.Value, &i.Type); err != nil {
if err := rows.Scan(&i.Queryid, &i.Type, &i.Value); err != nil {
return nil, err
}
items = append(items, &i)
@@ -133,67 +179,60 @@ func (q *Queries) ListQueryRequirementValues(ctx context.Context, arg *ListQuery
return items, nil
}
const listUnsyncedQueriesByDocId = `-- name: ListUnsyncedQueriesByDocId :many
WITH RECURSIVE unsyncedQueries AS (
SELECT dt.queryId, dt.requiredIds
from documents as d
JOIN fullActiveCollectors as c on d.jobId = c.jobId
JOIN collectorQueryDependencyTree as dt on c.id = dt.collectorId
const listUnsyncedNoDepsQueriesByDocId = `-- name: ListUnsyncedNoDepsQueriesByDocId :many
WITH docs as (
SELECT id, jobId from documents where id = $1
),
unsyncedQueries AS (
SELECT DISTINCT dt.queryId, dt.requiredIds, r.value, d.jobID, d.id
from docs as d
JOIN collectorQueryDependencyTree as dt on d.jobId = dt.jobId
JOIN fullActiveCollectors as c on c.id = dt.collectorId
LEFT JOIN results as r on r.queryId = dt.queryId
and r.documentId = d.id
and r.queryVersion = dt.queryVersion
and r.cleanVersion >= c.minCleanVersion and r.textVersion >= c.minTextVersion
where d.id = $1 and r.value is null
UNION
SELECT DISTINCT dt.queryId, dt.requiredIds
FROM unsyncedQueries as u
JOIN collectorQueryDependencyTree as dt ON u.queryId = any(dt.requiredIds)
where r.value is null
)
SELECT id, type, activeversion, latestversion, config, requiredids FROM fullActiveQueries
WHERE id in (SELECT queryId FROM unsyncedQueries)
SELECT DISTINCT queryId FROM unsyncedQueries as baseuq
WHERE NOT EXISTS (
SELECT 1 FROM unsyncedQueries as uq WHERE uq.queryId = any(baseuq.requiredIds)
)
`
// ListUnsyncedQueriesByDocId
// ListUnsyncedNoDepsQueriesByDocId
//
// WITH RECURSIVE unsyncedQueries AS (
// SELECT dt.queryId, dt.requiredIds
// from documents as d
// JOIN fullActiveCollectors as c on d.jobId = c.jobId
// JOIN collectorQueryDependencyTree as dt on c.id = dt.collectorId
// WITH docs as (
// SELECT id, jobId from documents where id = $1
// ),
// unsyncedQueries AS (
// SELECT DISTINCT dt.queryId, dt.requiredIds, r.value, d.jobID, d.id
// from docs as d
// JOIN collectorQueryDependencyTree as dt on d.jobId = dt.jobId
// JOIN fullActiveCollectors as c on c.id = dt.collectorId
// LEFT JOIN results as r on r.queryId = dt.queryId
// and r.documentId = d.id
// and r.queryVersion = dt.queryVersion
// and r.cleanVersion >= c.minCleanVersion and r.textVersion >= c.minTextVersion
// where d.id = $1 and r.value is null
//
// UNION
//
// SELECT DISTINCT dt.queryId, dt.requiredIds
// FROM unsyncedQueries as u
// JOIN collectorQueryDependencyTree as dt ON u.queryId = any(dt.requiredIds)
// where r.value is null
// )
// SELECT id, type, activeversion, latestversion, config, requiredids FROM fullActiveQueries
// WHERE id in (SELECT queryId FROM unsyncedQueries)
func (q *Queries) ListUnsyncedQueriesByDocId(ctx context.Context, id pgtype.UUID) ([]*Fullactivequery, error) {
rows, err := q.db.Query(ctx, listUnsyncedQueriesByDocId, id)
// SELECT DISTINCT queryId FROM unsyncedQueries as baseuq
// WHERE NOT EXISTS (
// SELECT 1 FROM unsyncedQueries as uq WHERE uq.queryId = any(baseuq.requiredIds)
// )
func (q *Queries) ListUnsyncedNoDepsQueriesByDocId(ctx context.Context, dollar_1 pgtype.UUID) ([]pgtype.UUID, error) {
rows, err := q.db.Query(ctx, listUnsyncedNoDepsQueriesByDocId, dollar_1)
if err != nil {
return nil, err
}
defer rows.Close()
items := []*Fullactivequery{}
items := []pgtype.UUID{}
for rows.Next() {
var i Fullactivequery
if err := rows.Scan(
&i.ID,
&i.Type,
&i.Activeversion,
&i.Latestversion,
&i.Config,
&i.Requiredids,
); err != nil {
var queryid pgtype.UUID
if err := rows.Scan(&queryid); err != nil {
return nil, err
}
items = append(items, &i)
items = append(items, queryid)
}
if err := rows.Err(); err != nil {
return nil, err
@@ -201,8 +240,8 @@ func (q *Queries) ListUnsyncedQueriesByDocId(ctx context.Context, id pgtype.UUID
return items, nil
}
const setResult = `-- name: SetResult :one
INSERT INTO results (queryId, documentId, value, cleanVersion, textVersion, queryVersion) VALUES ($1, $2, $3, $4, $5, $6) RETURNING id
const setResult = `-- name: SetResult :exec
INSERT INTO results (queryId, documentId, value, cleanVersion, textVersion, queryVersion) VALUES ($1, $2, $3, $4, $5, $6)
`
type SetResultParams struct {
@@ -216,9 +255,9 @@ type SetResultParams struct {
// SetResult
//
// INSERT INTO results (queryId, documentId, value, cleanVersion, textVersion, queryVersion) VALUES ($1, $2, $3, $4, $5, $6) RETURNING id
func (q *Queries) SetResult(ctx context.Context, arg *SetResultParams) (pgtype.UUID, error) {
row := q.db.QueryRow(ctx, setResult,
// INSERT INTO results (queryId, documentId, value, cleanVersion, textVersion, queryVersion) VALUES ($1, $2, $3, $4, $5, $6)
func (q *Queries) SetResult(ctx context.Context, arg *SetResultParams) error {
_, err := q.db.Exec(ctx, setResult,
arg.Queryid,
arg.Documentid,
arg.Value,
@@ -226,7 +265,5 @@ func (q *Queries) SetResult(ctx context.Context, arg *SetResultParams) (pgtype.U
arg.Textversion,
arg.Queryversion,
)
var id pgtype.UUID
err := row.Scan(&id)
return id, err
return err
}