1 files changed, 57 insertions, 30 deletions

diff --git a/src/imdbdata.go b/src/imdbdata.go
index c970c9b..30095ea 100644
--- a/src/imdbdata.go
+++ b/src/imdbdata.go
@@ -1,8 +1,8 @@
 package main
 
 import (
+	"bufio"
 	"compress/gzip"
-	"encoding/csv"
 	"fmt"
 	"io"
 	"log"
@@ -10,6 +10,7 @@ import (
 	"os"
 	"path/filepath"
 	"strconv"
+	"strings"
 )
 
 const (
@@ -24,6 +25,7 @@ func dataPath(name string) string {
 
 // downloadFile fetches url and writes to dst, overwriting if present.
 func downloadFile(dst, url string) error {
+	return nil
 	resp, err := http.Get(url)
 	if err != nil {
 		return fmt.Errorf("http get: %w", err)
@@ -122,8 +124,9 @@ type basicEntry struct {
 	RuntimeMinutes *int
 }
 
-// parseTSV reads a TSV file and calls fn for each row (after header).
+// parseTSV reads a TSV file line by line and calls fn for each row (after header).
 // Only rows where keep(tconst) is true are passed to fn.
+// Uses simple tab-splitting to avoid csv.Reader quote issues with large files.
 func parseTSV(path string, keep func(string) bool, fn func(record []string) error) error {
 	f, err := os.Open(path)
 	if err != nil {
@@ -131,39 +134,34 @@ func parseTSV(path string, keep func(string) bool, fn func(record []string) erro
 	}
 	defer f.Close()
 
-	r := csv.NewReader(f)
-	r.Comma = '\t'
-	r.LazyQuotes = true
-	r.FieldsPerRecord = -1
+	scanner := bufio.NewScanner(f)
 
 	// skip header
-	if _, err := r.Read(); err != nil {
-		return fmt.Errorf("read header: %w", err)
+	if !scanner.Scan() {
+		return fmt.Errorf("empty file")
 	}
 
-	for {
-		rec, err := r.Read()
-		if err == io.EOF {
-			break
-		}
-		if err != nil {
-			return fmt.Errorf("read row: %w", err)
+	for scanner.Scan() {
+		line := scanner.Text()
+		if line == "" {
+			continue
 		}
 
-		if len(rec) < 1 {
+		parts := strings.Split(line, "\t")
+		if len(parts) < 1 {
 			continue
 		}
 
-		tconst := rec[0]
+		tconst := parts[0]
 		if !keep(tconst) {
 			continue
 		}
 
-		if err := fn(rec); err != nil {
+		if err := fn(parts); err != nil {
 			return err
 		}
 	}
-	return nil
+	return scanner.Err()
 }
 
 // parseTitleRatings reads the ratings TSV and collects entries for the given ids.
@@ -198,7 +196,8 @@ func (a *App) parseTitleBasics(ids map[string]bool) (map[string]basicEntry, erro
 	err := parseTSV(dataPath("title.basics.tsv"),
 		func(s string) bool { return ids[s] },
 		func(rec []string) error {
-			if len(rec) < 7 {
+			// tconst, titleType, primaryTitle, originalTitle, isAdult, startYear, endYear, runtimeMinutes, genres
+			if len(rec) < 8 {
 				return nil
 			}
 
@@ -208,15 +207,17 @@ func (a *App) parseTitleBasics(ids map[string]bool) (map[string]basicEntry, erro
 				OriginalTitle: rec[3],
 			}
 
-			if rec[4] != "\\N" && rec[4] != "" {
-				v, err := strconv.Atoi(rec[4])
+			// startYear is field 5 (0-indexed)
+			if rec[5] != "\\N" && rec[5] != "" {
+				v, err := strconv.Atoi(rec[5])
 				if err == nil {
 					entry.StartYear = &v
 				}
 			}
 
-			if rec[5] != "\\N" && rec[5] != "" {
-				v, err := strconv.Atoi(rec[5])
+			// runtimeMinutes is field 7 (0-indexed)
+			if rec[7] != "\\N" && rec[7] != "" {
+				v, err := strconv.Atoi(rec[7])
 				if err == nil {
 					entry.RuntimeMinutes = &v
 				}
@@ -318,25 +319,32 @@ func (a *App) fetchAndUpdateImdbData() error {
 		return err
 	}
 
-	// Gather all imdb_ids needing updates
-	ids, err := a.getImdbIDsWithoutRating()
+	// Gather all imdb_ids needing rating updates
+	ratingIDs, err := a.getImdbIDsWithoutRating()
 	if err != nil {
 		return err
 	}
-	if len(ids) == 0 {
+	if len(ratingIDs) == 0 {
 		log.Println("fetchAndUpdateImdbData: no entries need updating")
 		return nil
 	}
-	log.Printf("fetchAndUpdateImdbData: %d entries to update", len(ids))
+	log.Printf("fetchAndUpdateImdbData: %d entries need rating update", len(ratingIDs))
+
+	// Gather all imdb_ids for basics update
+	allIDs, err := a.getAllImdbIDs()
+	if err != nil {
+		return err
+	}
+	log.Printf("fetchAndUpdateImdbData: %d entries total for basics update", len(allIDs))
 
 	// Parse datasets
-	ratings, err := a.parseTitleRatings(ids)
+	ratings, err := a.parseTitleRatings(ratingIDs)
 	if err != nil {
 		return fmt.Errorf("parseTitleRatings: %w", err)
 	}
 	log.Printf("fetchAndUpdateImdbData: found ratings for %d entries", len(ratings))
 
-	basics, err := a.parseTitleBasics(ids)
+	basics, err := a.parseTitleBasics(allIDs)
 	if err != nil {
 		return fmt.Errorf("parseTitleBasics: %w", err)
 	}
@@ -349,3 +357,22 @@ func (a *App) fetchAndUpdateImdbData() error {
 
 	return nil
 }
+
+// getAllImdbIDs returns all imdb_id values in the imdb table.
+func (a *App) getAllImdbIDs() (map[string]bool, error) {
+	rows, err := a.DB.Query(`SELECT imdb_id FROM imdb`)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+
+	ids := make(map[string]bool)
+	for rows.Next() {
+		var id string
+		if err := rows.Scan(&id); err != nil {
+			return nil, err
+		}
+		ids[id] = true
+	}
+	return ids, rows.Err()
+}