feat(import): Switching graphql API to Rest API for auto imports (#739)

This pr changes the Graphql API approach for the REST API on the auto
imports feature . This way we can import repos without a GitHub token
.There are some sync that won't work without proper validation, those
are :

1. PR Commits
2. Pr Reviews
3. Repo Issues
4. Repo metadata
5. Pull Requests
6. Repo Stars

In this content we also reintroduce stdout logging to the importer

To test this  you need to 

1. Log into mergestat
8. go to /repos page and click on  add repositories
9. Choose an user/org with public repos 
10. Those repos should be appearing  after a while

Closes #540

cmd/worker/main.go

@@ -5,13 +5,6 @@ import (
 	"database/sql"
 	"errors"
 	"fmt"
-	"github.com/mergestat/mergestat/internal/cron"
-	"github.com/mergestat/mergestat/internal/jobs/repo"
-	"github.com/mergestat/mergestat/internal/syncer"
-	"github.com/mergestat/mergestat/internal/timeout"
-	"github.com/mergestat/sqlq"
-	"github.com/mergestat/sqlq/runtime/embed"
-	"github.com/mergestat/sqlq/schema"
 	"net/http"
 	_ "net/http/pprof"
 	"net/url"
@@ -22,6 +15,14 @@ import (
 	"syscall"
 	"time"
 
+	"github.com/mergestat/mergestat/internal/cron"
+	"github.com/mergestat/mergestat/internal/jobs/repo"
+	"github.com/mergestat/mergestat/internal/syncer"
+	"github.com/mergestat/mergestat/internal/timeout"
+	"github.com/mergestat/sqlq"
+	"github.com/mergestat/sqlq/runtime/embed"
+	"github.com/mergestat/sqlq/schema"
+
 	"github.com/go-git/go-git/v5"
 	"github.com/golang-migrate/migrate/v4"
 	_ "github.com/golang-migrate/migrate/v4/database/postgres"
@@ -247,7 +248,7 @@ func main() {
 	var worker, _ = embed.NewWorker(upstream, embed.WorkerConfig{Queues: queues})
 
 	// register job handlers for types implemented by this worker
-	_ = worker.Register("repos/auto-import", repo.AutoImport(pool, db))
+	_ = worker.Register("repos/auto-import", repo.AutoImport(&logger, pool))
 
 	// TODO all of the following "params" should be configurable
 	// either via the database/app or possibly with env vars

internal/db/queries.sql

@@ -86,6 +86,9 @@ INSERT INTO mergestat.repo_sync_logs (log_type, message, repo_sync_queue_id) VAL
 -- name: SetSyncJobStatus :exec
 SELECT mergestat.set_sync_job_status(@Status::TEXT, @ID::BIGINT);
 
+-- name: FetchGitHubToken :one
+SELECT pgp_sym_decrypt(credentials, $1) FROM mergestat.service_auth_credentials WHERE type = 'GITHUB_PAT' ORDER BY created_at DESC LIMIT 1;
+
 -- We use a CTE here to retrieve all the repo_sync_jobs that were previously enqueued, to make sure that we *do not* re-enqueue anything new until the previously enqueued jobs are *completed*.
 -- This allows us to make sure all repo syncs complete before we reschedule a new batch.
 -- We have now also added a concept of type groups which allows us to apply this same logic but by each group type which is where the PARTITION BY clause comes into play