Convert HallOfBeornDump to CSV for ALEP pipeline

cmd/beornextract/BUILD

@@ -15,6 +15,7 @@ go_library(
     deps = [
         "//cmd/beornextract/types",
         "@com_github_grokify_html_strip_tags_go//:html-strip-tags-go",
+        "@com_github_jessevdk_go_flags//:go_default_library",
         "@io_bazel_rules_go//go/tools/bazel:go_default_library",
     ],
 )

cmd/beornextract/main.go

@@ -14,10 +14,24 @@ import (
 
 	"github.com/bazelbuild/rules_go/go/tools/bazel"
 	strip "github.com/grokify/html-strip-tags-go"
+	"github.com/jessevdk/go-flags"
 	"github.com/squk/lotr/cmd/beornextract/types"
 )
 
+type Options struct {
+	RawConversion bool `short:"r" long:"raw" description:"Enable to keep the original text from HallOfBeorn dump. Enable to prep for ALEP pipeline."`
+}
+
+var opts = Options{
+	RawConversion: false,
+}
+
 func main() {
+	_, err := flags.ParseArgs(&opts, os.Args)
+	if err != nil {
+		panic(err)
+	}
+
 	fmt.Println("LOTR CARD PARSE")
 	f, err := bazel.Runfile(".")
 	if err != nil {
@@ -36,8 +50,8 @@ func main() {
 	}
 
 	// Open our jsonFile
-	jsonFile, err := os.Open("cmd/beornextract/data/Bot.Cards.json")
+	// jsonFile, err := os.Open("cmd/beornextract/data/Bot.Cards.json")
-	// jsonFile, err := bazel.Runfile("cmd/beornextract/data/Bot.Cards.json")
+	jsonFile, err := os.Open("cmd/beornextract/data/Export.Cards.json")
 	// if we os.Open returns an error then handle it
 	if err != nil {
 		fmt.Println(err)
@@ -82,7 +96,7 @@ func main() {
 				card.TypeName,
 				card.SphereName,
 				card.Traits,
-				extractKeywords(card.Text),
+				findKeywords(card.Text),
 				card.Cost,
 				card.EngagementCost,
 				strconv.Itoa(card.Threat),
@@ -90,8 +104,8 @@ func main() {
 				strconv.Itoa(card.Attack),
 				strconv.Itoa(card.Defense),
 				strconv.Itoa(card.Health),
-				"", // Quest Points
+				card.QuestPoints,
-				"", // Victory Points
+				strconv.Itoa(card.VictoryPoints),
 				"", // Special Icon
 				transformText(card.Name, card.Text),
 				card.Flavor,
@@ -105,12 +119,18 @@ func main() {
 }
 
 func transformText(name, text string) string {
-	return strip.StripTags(
+	if opts.RawConversion {
-		strings.ReplaceAll(text, name, "[name]"), // insert name tag
+		return text
-	)
+	}
+
+	out := strings.ReplaceAll(text, name, "[name]") // insert name tag
+	out = strip.StripTags(out)
+	out = keywordPattern.ReplaceAllLiteralString(out, "")
+	return strings.TrimSpace(out)
 }
 
-func extractKeywords(text string) string {
+var keywordPattern = regexp.MustCompile(`((?:(?:[A-Z][a-z]+(\.|\s[0-9]+\.)\s*)+))`)
-	pattern := regexp.MustCompile(`((?:(?:[A-Z][a-z]+(\.|\s[0-9]+\.)\s)+))`)
+
-	return strings.TrimSpace(pattern.FindString(text))
+func findKeywords(text string) string {
+	return strings.TrimSpace(keywordPattern.FindString(text))
 }

cmd/beornextract/types/card.go

@@ -9,29 +9,19 @@ type HallOfBeornCard struct {
 	SphereCode    string `json:"sphere_code,omitempty"`
 	SphereName    string `json:"sphere_name,omitempty"`
 	Position      int    `json:"position,omitempty"`
-	Code        string `json:"code,omitempty"`
-	Cost        string `json:"cost,omitempty"`
-	Name        string `json:"name,omitempty"`
-	Traits      string `json:"traits,omitempty"`
-	Text        string `json:"text,omitempty"`
-	Flavor      string `json:"flavor,omitempty"`
-	IsUnique    bool   `json:"is_unique,omitempty"`
 	Threat        int    `json:"threat,omitempty"`
 	Willpower     int    `json:"willpower,omitempty"`
 	Attack        int    `json:"attack,omitempty"`
 	Defense       int    `json:"defense,omitempty"`
 	Health        int    `json:"health,omitempty"`
-	Quantity    int    `json:"quantity,omitempty"`
-	DeckLimit   int    `json:"deck_limit,omitempty"`
-	Illustrator string `json:"illustrator,omitempty"`
 	Octgnid       string `json:"octgnid,omitempty"`
 	HasErrata     bool   `json:"has_errata,omitempty"`
 	URL           string `json:"url,omitempty"`
+	QuestPoints   string `json:"quest,omitempty"`
+	VictoryPoints int    `json:"victory,omitempty"`
 	Imagesrc      string `json:"imagesrc,omitempty"`
 
 	EncounterSet   string `json:"encounter_set,omitempty"`
 	EngagementCost string `json:"engagement_cost,omitempty"`
 	ThreatStrength int    `json:"threat_strength,omitempty"`
 }
-
-type NormalizedCard struct{}

deps.bzl

@@ -37,6 +37,13 @@ def go_dependencies():
         sum = "h1:0fThFwLbW7P/kOiTBs03FsJSV9RM2M/Q/MOnCQxKMo0=",
         version = "v0.0.1",
     )
+    go_repository(
+        name = "com_github_jessevdk_go_flags",
+        build_file_proto_mode = "disable_global",
+        importpath = "github.com/jessevdk/go-flags",
+        sum = "h1:1jKYvbxEjfUl0fmqTCOfonvskHHXMjBySTLW4y9LFvc=",
+        version = "v1.5.0",
+    )
 
     go_repository(
         name = "com_github_mattn_go_sqlite3",

go.mod

@@ -7,5 +7,7 @@ require github.com/davecgh/go-spew v1.1.1
 require (
 	github.com/bazelbuild/rules_go v0.41.0 // indirect
 	github.com/grokify/html-strip-tags-go v0.0.1 // indirect
+	github.com/jessevdk/go-flags v1.5.0 // indirect
 	github.com/mattn/go-sqlite3 v1.14.17 // indirect
+	golang.org/x/sys v0.0.0-20210510120138-977fb7262007 // indirect
 )

go.sum

@@ -4,5 +4,10 @@ github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/grokify/html-strip-tags-go v0.0.1 h1:0fThFwLbW7P/kOiTBs03FsJSV9RM2M/Q/MOnCQxKMo0=
 github.com/grokify/html-strip-tags-go v0.0.1/go.mod h1:2Su6romC5/1VXOQMaWL2yb618ARB8iVo6/DR99A6d78=
+github.com/jessevdk/go-flags v1.5.0 h1:1jKYvbxEjfUl0fmqTCOfonvskHHXMjBySTLW4y9LFvc=
+github.com/jessevdk/go-flags v1.5.0/go.mod h1:Fw0T6WPc1dYxT4mKEZRfG5kJhaTDP9pj1c2EWnYs/m4=
 github.com/mattn/go-sqlite3 v1.14.17 h1:mCRHCLDUBXgpKAqIKsaAaAsrAlbkeomtRFKXh2L6YIM=
 github.com/mattn/go-sqlite3 v1.14.17/go.mod h1:2eHXhiwb8IkHr+BDWZGa96P6+rkvnG63S2DGjv9HUNg=
+golang.org/x/sys v0.0.0-20210320140829-1e4c9ba3b0c4/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+golang.org/x/sys v0.0.0-20210510120138-977fb7262007 h1:gG67DSER+11cZvqIMb8S8bt0vZtiN6xWYARwirrOSfE=
+golang.org/x/sys v0.0.0-20210510120138-977fb7262007/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=