// Command import-candidates sources cleared IT and cyber leads from People Data Labs (the Galapagos
// preset) and upserts them into the tenant candidates table tagged external, writing a compliance row
// per record (source, provider, query, legal basis, collected at). Clearance is left unknown: absence
// is never a negative. Run with `make import-candidates` (small free tier batches). Flags let you set
// the batch size, include contact fields, and set the legal basis. Reads PDL_API_KEY plus
// MM_MIGRATOR_DATABASE_URL plus optional MM_DEMO_ORG_ID from env.
package main

import (
	"context"
	"encoding/json"
	"errors"
	"flag"
	"fmt"
	"log"
	"os"
	"strings"

	"github.com/google/uuid"
	"github.com/jackc/pgx/v5/pgxpool"

	"github.com/missionworkforce/missionmatch/internal/platform/database"
	"github.com/missionworkforce/missionmatch/internal/platform/database/dbgen"
	"github.com/missionworkforce/missionmatch/internal/platform/extraction"
	"github.com/missionworkforce/missionmatch/internal/platform/sourcing"
)

const defaultOrg = "11111111-1111-1111-1111-111111111111"

func main() {
	size := flag.Int("size", 10, "number of leads to source (keep small for the free tier)")
	includeContact := flag.Bool("include-contact", false, "store email and phone (off by default)")
	legalBasis := flag.String("legal-basis",
		"Legitimate interest: recruitment sourcing of publicly available professional data",
		"the lawful basis recorded in the compliance log")
	flag.Parse()

	ctx := context.Background()
	key := strings.TrimSpace(os.Getenv("PDL_API_KEY"))
	dsn := firstNonEmpty(os.Getenv("MM_MIGRATOR_DATABASE_URL"), os.Getenv("MM_DATABASE_URL"))
	orgStr := firstNonEmpty(os.Getenv("MM_DEMO_ORG_ID"), defaultOrg)
	if key == "" {
		log.Fatal("import-candidates: PDL_API_KEY not set (server side env only)")
	}
	if dsn == "" {
		log.Fatal("import-candidates: set MM_MIGRATOR_DATABASE_URL (or MM_DATABASE_URL)")
	}
	orgID, err := uuid.Parse(orgStr)
	if err != nil {
		log.Fatalf("import-candidates: bad org id %q: %v", orgStr, err)
	}

	pool, err := pgxpool.New(ctx, dsn)
	if err != nil {
		log.Fatalf("import-candidates: db connect: %v", err)
	}
	defer pool.Close()
	if err := pool.Ping(ctx); err != nil {
		log.Fatalf("import-candidates: db ping: %v", err)
	}
	// Writes the candidate and its compliance row with an explicit organization_id and no
	// app.current_org_id, so the BYPASSRLS migrator role is required or RLS rejects every insert.
	if err := database.AssertMigratorRole(ctx, pool); err != nil {
		log.Fatalf("import-candidates: %v", err)
	}
	q := dbgen.New(pool)
	ext := extraction.NewDeterministic() // certs and agencies from the profile text; skills come structured

	prov := sourcing.NewPDL(key, os.Getenv("PDL_API_BASE"), *includeContact)
	queryObj := sourcing.GalapagosSearch()
	queryJSON, _ := json.Marshal(queryObj)

	log.Printf("import-candidates: provider=%q size=%d include_contact=%v org=%s", prov.Name(), *size, *includeContact, orgID)

	people, err := prov.Search(ctx, sourcing.SearchQuery{ESQuery: queryObj, Size: *size})
	if err != nil {
		if errors.Is(err, sourcing.ErrThrottled) {
			log.Printf("import-candidates: provider rate limited (429), stopping run gracefully")
			return
		}
		log.Fatalf("import-candidates: search: %v", err)
	}
	log.Printf("import-candidates: provider returned %d leads", len(people))

	upserted := 0
	for _, sp := range people {
		if err := importOne(ctx, pool, q, ext, orgID, sp, prov.Name(), string(queryJSON), *legalBasis); err != nil {
			log.Printf("import-candidates: %s: %v", sp.ExternalID, err)
			continue
		}
		upserted++
	}
	log.Printf("import-candidates: DONE, upserted %d external candidates with compliance rows", upserted)
}

// importOne resolves and upserts one sourced person (identity resolution + structured skills, certs, and
// agencies) and writes its compliance row inside a single transaction, so a lead is never persisted
// without the legal basis record that justifies holding it. Either both commit or neither does, which
// keeps the sourcing provenance promise (a compliance row per record) true even if the second write
// fails. Only a fully committed record counts toward the total.
func importOne(ctx context.Context, pool *pgxpool.Pool, base *dbgen.Queries, ext extraction.Extractor, orgID uuid.UUID, sp sourcing.SourcedPerson, provider, queryJSON, legalBasis string) error {
	tx, err := pool.Begin(ctx)
	if err != nil {
		return fmt.Errorf("begin tx: %w", err)
	}
	defer func() { _ = tx.Rollback(ctx) }() // no op after a successful Commit

	q := base.WithTx(tx)
	candID, err := sourcing.IngestPerson(ctx, q, orgID, sp, ext)
	if err != nil {
		return err
	}
	if err := q.InsertSourcingLog(ctx, dbgen.InsertSourcingLogParams{
		OrganizationID: orgID,
		CandidateID:    &candID,
		ExternalID:     sp.ExternalID,
		Source:         "PDL",
		Provider:       provider,
		Query:          queryJSON,
		LegalBasis:     legalBasis,
	}); err != nil {
		return fmt.Errorf("insert sourcing log: %w", err)
	}
	if err := tx.Commit(ctx); err != nil {
		return fmt.Errorf("commit tx: %w", err)
	}
	return nil
}

func firstNonEmpty(vals ...string) string {
	for _, v := range vals {
		if s := strings.TrimSpace(v); s != "" {
			return s
		}
	}
	return ""
}