feat: ✨ update amazon metadata

data/amazon/stats/category_list.json

@@ -0,0 +1,1425 @@
+[
+    "jerseys",
+    "baseball caps",
+    "running",
+    "gun holsters",
+    "imported",
+    "equestrian sports",
+    "auto accessories",
+    "skiing",
+    "dcor",
+    "swimming",
+    "pants",
+    "protective gear",
+    "hunting knives",
+    "shirts",
+    "martial arts",
+    "jackets",
+    "skateboarding",
+    "shorts",
+    "yoga",
+    "archery",
+    "socks",
+    "airsoft",
+    "cell phone accessories",
+    "sports water bottles",
+    "camp kitchen",
+    "sport watches",
+    "optics",
+    "gloves",
+    "sports souvenirs",
+    "soccer",
+    "strength training equipment",
+    "backpacking packs",
+    "swimwear",
+    "sports equipment",
+    "grips",
+    "paintball",
+    "tennis",
+    "machine wash",
+    "skateboard parts",
+    "outdoor flags",
+    "gun maintenance",
+    "tack",
+    "weapons",
+    "snowboarding",
+    "artwork",
+    "golf clubs",
+    "goggles",
+    "boys",
+    "training equipment",
+    "baby clothing",
+    "climbing",
+    "tactical knives",
+    "gun scopes",
+    "terminal tackle",
+    "hiking daypacks",
+    "dance",
+    "girls",
+    "bedding",
+    "hunting apparel",
+    "other team sports",
+    "fixed blade hunting knives",
+    "fitness technology",
+    "winter sports accessories",
+    "decals",
+    "hard lures",
+    "hats",
+    "hunting accessories",
+    "polo shirts",
+    "on-course accessories",
+    "coolers",
+    "rifle scopes",
+    "gunsmithing tools",
+    "license plate frames",
+    "football",
+    "wristbands",
+    "camping furniture",
+    "tires",
+    "swords",
+    "diving suits",
+    "bikes",
+    "toy figures",
+    "boxing",
+    "wheels",
+    "fishing apparel",
+    "rods",
+    "disc sports",
+    "ornaments",
+    "basketball",
+    "casino equipment",
+    "drivetrain components",
+    "key chains",
+    "pet gear",
+    "reels",
+    "bats",
+    "balls",
+    "sunglasses",
+    "kayaking",
+    "surfing",
+    "helmets",
+    "folding hunting knives",
+    "footwear",
+    "folding knives",
+    "bowling",
+    "electrical equipment",
+    "personal defense equipment",
+    "throw blankets",
+    "mitts",
+    "personal care",
+    "head covers",
+    "polyester",
+    "wallets",
+    "earrings",
+    "player equipment",
+    "wetsuits",
+    "saddles",
+    "tents",
+    "boat engine parts",
+    "exercise bands",
+    "ice hockey",
+    "cotton",
+    "sleeping bags",
+    "golf balls",
+    "compression shorts",
+    "golf cart accessories",
+    "cardio training",
+    "adult helmets",
+    "fishing line",
+    "racquets",
+    "disc golf",
+    "mats",
+    "brake parts",
+    "golf equipment",
+    "artificial bait",
+    "headbands",
+    "waist packs",
+    "one-piece suits",
+    "equipment bags",
+    "sports sunglasses",
+    "baseball mitts",
+    "paracord",
+    "decks",
+    "golf club bags",
+    "lacrosse",
+    "standard skateboards",
+    "hooks",
+    "office products",
+    "gun scope mounts",
+    "toss games",
+    "drivers",
+    "gun sights",
+    "discs",
+    "fitness trackers",
+    "travel mugs",
+    "hardware",
+    "coffee mugs",
+    "boat covers",
+    "baseball bats",
+    "hydration packs",
+    "compression",
+    "scooters",
+    "towels",
+    "bracelets",
+    "chairs",
+    "sports medicine",
+    "hand wash",
+    "game calls",
+    "furniture",
+    "exercise videos",
+    "electronics",
+    "backpacks",
+    "officially licensed",
+    "handlebars",
+    "sweaters",
+    "tactical backpacks",
+    "bath",
+    "riding boots",
+    "photos",
+    "machine washable",
+    "lanterns",
+    "pedals",
+    "volleyball",
+    "leggings",
+    "leotards",
+    "game handling",
+    "rifles",
+    "gun belts",
+    "table tennis",
+    "trading cards",
+    "china",
+    "wall decals",
+    "football equipment",
+    "topwater lures",
+    "golf club parts",
+    "compasses",
+    "nets",
+    "longboards",
+    "putters",
+    "bows",
+    "tubes",
+    "ball markers",
+    "headlights",
+    "hydration",
+    "slippers",
+    "pet clothing",
+    "jigs",
+    "infielder's mitts",
+    "quartz movement",
+    "boxing gloves",
+    "swim caps",
+    "racquet accessories",
+    "bike locks",
+    "poker equipment",
+    "freeze-dried food",
+    "heart rate monitors",
+    "darts",
+    "tent accessories",
+    "pakistan",
+    "jump ropes",
+    "insect repellent",
+    "boat motors",
+    "cornhole",
+    "honduras",
+    "pistols",
+    "rifle cases",
+    "street signs",
+    "bike pumps",
+    "tactical flashlights",
+    "tactical pouches",
+    "boat cabin products",
+    "tackle storage",
+    "area rugs",
+    "downhill skiing",
+    "kids' bikes",
+    "fishing hats",
+    "gymnastics",
+    "boots",
+    "tights",
+    "ice skating",
+    "balance trainers",
+    "exercise balls",
+    "compression tops",
+    "zipper closure",
+    "soft lures",
+    "hammocks",
+    "stove accessories",
+    "dry bags",
+    "shin guards",
+    "reflective gear",
+    "training gloves",
+    "boat plumbing",
+    "cue sticks",
+    "pet collars",
+    "boat trailer accessories",
+    "diving masks",
+    "mineral crystal",
+    "backpack accessories",
+    "gun slings",
+    "caps",
+    "novelty headwear",
+    "ice fishing",
+    "camping shelters",
+    "trekking poles",
+    "pennants",
+    "shells",
+    "family camping tents",
+    "halters",
+    "masks",
+    "fleece",
+    "two-piece suits",
+    "cheerleading",
+    "safety glasses",
+    "saddle pads",
+    "stand-up paddleboarding",
+    "made in usa",
+    "soccer equipment",
+    "stems",
+    "derailleurs",
+    "batting gloves",
+    "fly fishing",
+    "scarves",
+    "floormats",
+    "bangladesh",
+    "bike bells",
+    "air guns",
+    "standard balls",
+    "officially licensed by the nfl",
+    "bits",
+    "cart bags",
+    "camouflage accessories",
+    "complete sets",
+    "beach towels",
+    "gun cleaning kits",
+    "grill accessories",
+    "punching bags",
+    "snowmobiling",
+    "football gear",
+    "bike racks",
+    "water bottle cages",
+    "tees",
+    "badminton",
+    "maintenance supplies",
+    "exercise bikes",
+    "plaques",
+    "bowling balls",
+    "seat posts",
+    "trucks",
+    "pedometers",
+    "umbrellas",
+    "wall clocks",
+    "broadheads",
+    "shafts",
+    "internal frame backpacks",
+    "toy vehicles",
+    "sailing",
+    "tumblers",
+    "license plate covers",
+    "knives",
+    "backpacking tents",
+    "basketballs",
+    "irons",
+    "bingo equipment",
+    "goalkeeper gloves",
+    "knee pads",
+    "sun shelters",
+    "fixed-blade knives",
+    "wall banners",
+    "tactical bag accessories",
+    "cranksets",
+    "snowshoeing",
+    "sights",
+    "cricket",
+    "headlamps",
+    "rash guards",
+    "bindings",
+    "straps",
+    "car flags",
+    "tables",
+    "kayaks",
+    "shop tools",
+    "home gyms",
+    "triathlon",
+    "whistles",
+    "golf gear",
+    "braided line",
+    "hybrid clubs",
+    "car magnets",
+    "bags",
+    "footballs",
+    "dumbbells",
+    "visors",
+    "hunting arrows",
+    "ties",
+    "kick scooters",
+    "pepper spray",
+    "optics mounts",
+    "button-down shirts",
+    "precise swiss-quartz movement",
+    "flying discs",
+    "protective padding",
+    "treadmill belts",
+    "roller skates",
+    "sets",
+    "wakeboarding",
+    "can coolers",
+    "synthetic",
+    "clothing sets",
+    "snorkeling packages",
+    "belts",
+    "mountain bikes",
+    "basketball equipment",
+    "taillights",
+    "camping cookware",
+    "tools",
+    "hand strengtheners",
+    "hunting hats",
+    "softball mitts",
+    "spotting scopes",
+    "made in us",
+    "sledding",
+    "air mattresses",
+    "breeches",
+    "foam rollers",
+    "inline skates",
+    "targets",
+    "brake pads",
+    "indonesia",
+    "sleeping bag accessories",
+    "exercise mats",
+    "flights",
+    "baseball equipment",
+    "flies",
+    "racquet strings",
+    "duffle bags",
+    "dresses",
+    "cornhole sets",
+    "beer glasses",
+    "baseballs",
+    "bibs",
+    "skirts",
+    "mini helmets",
+    "chainrings",
+    "acrylic",
+    "protective body equipment",
+    "sticks",
+    "folding chairs",
+    "brakes",
+    "cable locks",
+    "sleeping pads",
+    "pins",
+    "snowshoes",
+    "fly tying materials",
+    "cruiser bikes",
+    "bobbleheads",
+    "vests",
+    "air rifles",
+    "dry boxes",
+    "tactical vests",
+    "hunting bags",
+    "diving fins",
+    "blankets",
+    "propellers",
+    "mouthguards",
+    "kids' helmets",
+    "slow-pitch softball bats",
+    "brake levers",
+    "wrestling",
+    "stand bags",
+    "classic fit",
+    "hand warmers",
+    "fairway woods",
+    "charms",
+    "waist trimmers",
+    "monofilament line",
+    "decoys",
+    "rings",
+    "entry mats",
+    "deck hardware",
+    "poker chips",
+    "harnesses",
+    "paddles",
+    "pistol cases",
+    "headgear",
+    "cleaners",
+    "games",
+    "trailer hitch covers",
+    "racquetball",
+    "tops",
+    "fins",
+    "lighters",
+    "magazines",
+    "fire starters",
+    "rugby",
+    "basketball court equipment",
+    "rod holders",
+    "board bags",
+    "blinds",
+    "briefs",
+    "jammers",
+    "navigation lights",
+    "bowstrings",
+    "forks",
+    "skinsuits",
+    "outfielder's mitts",
+    "cycling computers",
+    "elliptical trainers",
+    "kayak accessories",
+    "bike baskets",
+    "benches",
+    "headlight-taillight combinations",
+    "handlebar bags",
+    "headphones",
+    "fletches",
+    "treadmills",
+    "grip tape",
+    "hubs",
+    "first aid kits",
+    "bed-in-a-bag",
+    "stand-up paddleboards",
+    "rubber sole",
+    "made in the usa and imported",
+    "fenders",
+    "snowboards",
+    "bbs",
+    "outboard motors",
+    "quivers",
+    "floor pumps",
+    "outdoor statues",
+    "pet jerseys",
+    "trampolines",
+    "importado",
+    "racquet grips",
+    "wraps",
+    "inline skate parts",
+    "one size fits most",
+    "tote bags",
+    "swing trainers",
+    "sports water bottle accessories",
+    "rope",
+    "cuff links",
+    "pet t-shirts",
+    "bait rigs",
+    "full suits",
+    "squash",
+    "bars",
+    "coasters",
+    "hockey equipment",
+    "night vision",
+    "chains",
+    "full sized helmets",
+    "kettlebells",
+    "sauna suits",
+    "towables",
+    "seating",
+    "trolling motors",
+    "hand wraps",
+    "cargo racks",
+    "polarized",
+    "weight belts",
+    "nylon",
+    "usa",
+    "mat bags",
+    "kayak hardware",
+    "skis",
+    "scents",
+    "armwarmers",
+    "midrange discs",
+    "bearings",
+    "snow sleds",
+    "nocks",
+    "display cases",
+    "medicine balls",
+    "locking carabiners",
+    "foosball",
+    "tent stakes",
+    "lamps",
+    "balance boards",
+    "anchors",
+    "markers",
+    "seat packs",
+    "goals",
+    "foot gear",
+    "fly line",
+    "rackets",
+    "tackle boxes",
+    "spoons",
+    "sandals",
+    "batting helmets",
+    "ice skates",
+    "pet dresses",
+    "cooler accessories",
+    "bilge pumps",
+    "el salvador",
+    "roller skate parts",
+    "road bikes",
+    "pilates",
+    "valve caps",
+    "boats",
+    "standard flights",
+    "soccer field equipment",
+    "bike pack accessories",
+    "fast-pitch softball bats",
+    "rocker switches",
+    "gun stocks",
+    "lunch boxes",
+    "made in usa or imported",
+    "hunting dog equipment",
+    "handlebar tape",
+    "screen printed graphics",
+    "catcher's mitts",
+    "golf club bag accessories",
+    "leather",
+    "golf club head covers",
+    "complete paintball sets",
+    "canoeing",
+    "fly tying equipment",
+    "poles",
+    "open fire cookware",
+    "elastic closure",
+    "softballs",
+    "team logo embroidered on left chest",
+    "bottom brackets",
+    "stabilizers",
+    "magnets",
+    "pull-up bars",
+    "dry flies",
+    "sneakers",
+    "court equipment",
+    "recurve bows",
+    "webbing",
+    "tree stands",
+    "field equipment",
+    "bowling roller bags",
+    "field hockey",
+    "rigging",
+    "trailers",
+    "compound bows",
+    "pull on closure",
+    "girths",
+    "diving lights",
+    "frame-mounted pumps",
+    "diploma frames",
+    "earmuffs",
+    "turkey",
+    "pickleball",
+    "fishing kayaks",
+    "golf umbrellas",
+    "quality japanese-quartz movement",
+    "release aids",
+    "one size fits all",
+    "bmx",
+    "climbing holds",
+    "snorkels",
+    "bodyboards",
+    "aprons",
+    "reins",
+    "stun guns",
+    "steering equipment",
+    "exercise machine attachments",
+    "fencing",
+    "earplugs",
+    "men's",
+    "bait traps",
+    "cases",
+    "cooking utensils",
+    "luggage tags",
+    "magazine loaders",
+    "basketball gear",
+    "medals",
+    "throw pillows",
+    "sleepwear",
+    "tree stand accessories",
+    "swivels",
+    "bike mirrors",
+    "barrels",
+    "air gun pellets",
+    "air hockey",
+    "computer mouse pads",
+    "kids' tricycles",
+    "golf carts",
+    "headsets",
+    "photomints",
+    "bow maintenance accessories",
+    "footprints",
+    "bumper stickers",
+    "electric bicycles",
+    "blocks",
+    "skate bags",
+    "heavy bags",
+    "cornhole bags",
+    "ski bags",
+    "baseball gear",
+    "volleyballs",
+    "stopwatches",
+    "slingshots",
+    "made by reebok",
+    "children's inline skates",
+    "weights",
+    "targeting arrows",
+    "short sleeve",
+    "stirrups",
+    "gun brushes",
+    "diving gloves",
+    "bath accessories",
+    "picture frames",
+    "handcuffs",
+    "plugs",
+    "bow cases",
+    "cots",
+    "flatware",
+    "camping pillows",
+    "non-polarized",
+    "swiss quartz movement",
+    "shot glasses",
+    "elbow pads",
+    "leashes",
+    "tool kits",
+    "tank tops",
+    "plates",
+    "divot tools",
+    "skates",
+    "dice",
+    "resistance trainers",
+    "drawstring bags",
+    "drawstring closure",
+    "poly-cotton",
+    "plastic frame",
+    "brand new",
+    "recoil pads",
+    "seat clamps",
+    "suspension",
+    "shoulder pads",
+    "comforters",
+    "optics accessories",
+    "shifters",
+    "weight racks",
+    "interior lights",
+    "pet bandanas",
+    "nonlocking carabiners",
+    "figure skates",
+    "water filters",
+    "hybrid bikes",
+    "playground balls",
+    "rod racks",
+    "inflatable rafts",
+    "points",
+    "billiard balls",
+    "bag gloves",
+    "throwing equipment",
+    "cross-country skiing",
+    "attractants",
+    "rims",
+    "replacement wheels",
+    "bat accessories",
+    "fuel filters",
+    "ankle weights",
+    "crossbows",
+    "ninja weapons",
+    "stuff sacks",
+    "officially licensed by the nba",
+    "ice spearing equipment",
+    "bike covers",
+    "spurs",
+    "u-locks",
+    "kids' accessories",
+    "emergency fire starters",
+    "the north face size chart",
+    "self-inflating pads",
+    "tailgating toss games",
+    "golf bag towels",
+    "fight gloves",
+    "weight vests",
+    "shift levers",
+    "made in the usa",
+    "suspension forks",
+    "emblems",
+    "plastic lens",
+    "adjustable benches",
+    "officially licensed by the nhl",
+    "moisture wicking",
+    "optics rings",
+    "folding bikes",
+    "hockey gear",
+    "regulators",
+    "training fins",
+    "ladders",
+    "roller hockey",
+    "inversion equipment",
+    "canteens",
+    "emergency blankets",
+    "car rack accessories",
+    "steering wheel covers",
+    "bocce",
+    "cue chalk",
+    "fuel tanks",
+    "arm guards",
+    "lasers",
+    "hand wipe with damp cloth",
+    "officially licensed by the ncaa",
+    "heads",
+    "multifunction tools",
+    "batteries",
+    "shorty suits",
+    "predator",
+    "compression socks",
+    "front derailleurs",
+    "topographic maps",
+    "rests",
+    "regulator accessories",
+    "decoy accessories",
+    "shoe covers",
+    "singlets",
+    "poms",
+    "pillowcases",
+    "shuttlecocks",
+    "riding gloves",
+    "shovels",
+    "underwear",
+    "indoor bike storage",
+    "fluorocarbon line",
+    "paintballs",
+    "battery chargers",
+    "badge holders",
+    "paddle accessories",
+    "laser rangefinders",
+    "punching bag accessories",
+    "foosball accessories",
+    "bingo cards",
+    "deer",
+    "poker sets",
+    "seat covers",
+    "boat fenders",
+    "distance balls",
+    "canopies",
+    "officially licensed merchandise",
+    "vietnam",
+    "dry clean only",
+    "carry bags",
+    "children's roller skates",
+    "swim vests",
+    "duck",
+    "whips",
+    "pens",
+    "freestyle boards",
+    "open-hem sleeves",
+    "safety gear",
+    "bow slings",
+    "stunt scooters",
+    "aquatic fitness equipment",
+    "painting supplies",
+    "screen print graphics",
+    "thailand",
+    "tire covers",
+    "catcher chest protectors",
+    "fitted",
+    "fishing gloves",
+    "boot bags",
+    "made in usa and imported",
+    "taiwan",
+    "scent eliminators",
+    "gun cases",
+    "bingo daubers",
+    "bike frames",
+    "belt displays",
+    "wet flies",
+    "pet leashes",
+    "water bottles",
+    "lines",
+    "headstalls",
+    "luggage",
+    "tanks",
+    "practice nets",
+    "skate accessories",
+    "hitting nets",
+    "wax",
+    "wakeboards",
+    "pro boxing gloves",
+    "starter sets",
+    "flags",
+    "gun lubrication",
+    "shower curtains",
+    "football field equipment",
+    "pushup stands",
+    "backcountry equipment",
+    "tablet accessories",
+    "rangefinders",
+    "fixed gear bikes",
+    "tent tarps",
+    "bathroom towels",
+    "unitards",
+    "chalk bags",
+    "youth sizes 8-20",
+    "original packaging",
+    "korea",
+    "dartboards",
+    "leaders",
+    "seating accessories",
+    "foosball tables",
+    "cue racks",
+    "screenprint decoration",
+    "shooter sleeves",
+    "plumbing fittings",
+    "swaziland",
+    "chain locks",
+    "line spooling accessories",
+    "helmet accessories",
+    "alarm clocks",
+    "food processing",
+    "foot warmers",
+    "speed skates",
+    "mp3 player accessories",
+    "loaders",
+    "reflectors",
+    "bowls",
+    "rib knit collar",
+    "muff pocket",
+    "hunting knife sharpeners",
+    "marker maintenance equipment",
+    "bowling totes",
+    "embroidered logos",
+    "putting mats",
+    "lob wedges",
+    "shuffleboard",
+    "sleeve patches",
+    "made by outerstuff",
+    "lap counters",
+    "durable",
+    "quartz movement with digital display",
+    "combination tables",
+    "ball polishers",
+    "nicaragua",
+    "game feeders",
+    "pong games",
+    "air pistols",
+    "machine washable, line dry, no bleach",
+    "liners",
+    "protective flooring",
+    "practice balls",
+    "replacement parts",
+    "stomp pads",
+    "embroidered graphics",
+    "made by adidas",
+    "sand wedges",
+    "boresighters",
+    "kickboards",
+    "scratch resistant, shatter resistant sapphimax crystal",
+    "billiard cloth",
+    "co2 pumps",
+    "flasks",
+    "girdles",
+    "tumbling mats",
+    "polyester lining",
+    "brake calipers",
+    "bmx bikes",
+    "buoyancy compensators",
+    "fan packs",
+    "rubber",
+    "windbreakers",
+    "bushings",
+    "japanese quartz movement",
+    "croquet",
+    "cables",
+    "sheets",
+    "seats",
+    "collars",
+    "officially licensed by mlb",
+    "weapon stands",
+    "goalposts",
+    "travel covers",
+    "leads",
+    "comfort bikes",
+    "traction cleats",
+    "table covers",
+    "slow-pitch softballs",
+    "swiss chronograph movement",
+    "bimini tops",
+    "locks",
+    "handgun scopes",
+    "cue tips",
+    "workstands",
+    "field dressing accessories",
+    "waste baskets",
+    "gaiters",
+    "vibration dampeners",
+    "reservoirs",
+    "youth bow sets",
+    "tracksuits",
+    "groin protectors",
+    "wrist weights",
+    "child carrier trailers",
+    "ball retrievers",
+    "date function",
+    "ballet equipment",
+    "v-neck tee",
+    "rowers",
+    "surfboards",
+    "made in china",
+    "ring toss",
+    "pack covers",
+    "slim fit",
+    "backboard components",
+    "bait storage",
+    "pitching machines",
+    "bike lube",
+    "whitewater kayaks",
+    "snow tubes",
+    "gauges",
+    "spotlights",
+    "bike horns",
+    "dive skins",
+    "synthetic sole",
+    "practice swords",
+    "holsters",
+    "trophy mounts",
+    "disc brake sets",
+    "showers",
+    "bed pillows",
+    "night vision monoculars",
+    "officially licensed nfl product",
+    "officially licensed by major league baseball",
+    "step platforms",
+    "cleaning tools",
+    "crew neck jersey tee",
+    "compression sleeves",
+    "batting trainers",
+    "tank accessories",
+    "buckles",
+    "mexico",
+    "batting tees",
+    "step machines",
+    "horns",
+    "accessory kits",
+    "downriggers",
+    "made in the usa or imported",
+    "diving hoods",
+    "chalk",
+    "ultimate",
+    "chest protectors",
+    "punch mitts",
+    "rotors",
+    "hang dry",
+    "altimeters",
+    "loader accessories",
+    "screen print decoration",
+    "stools",
+    "leg machines",
+    "bridles",
+    "shotguns",
+    "laptop bags",
+    "ball racks",
+    "grip repair kits",
+    "vintage feel",
+    "saws",
+    "hangers",
+    "water purifiers",
+    "bear protection",
+    "kids' protective gear",
+    "ice augers",
+    "diving electronics",
+    "button closure",
+    "reel care accessories",
+    "protective mineral crystal dial window",
+    "barstools",
+    "adult",
+    "headset spacers",
+    "speedometers",
+    "shoe bags",
+    "cabinets",
+    "rubbers",
+    "leader rigging",
+    "hockey pucks",
+    "vibration platform machines",
+    "racing skates",
+    "gun solvents",
+    "camping grills",
+    "desk caddies",
+    "hard hats",
+    "diving boots",
+    "first baseman's mitts",
+    "arm machines",
+    "game table accessories",
+    "slacklines",
+    "battery switches",
+    "screenprint",
+    "bodysuits",
+    "single clubs",
+    "shore power cords",
+    "computer mice",
+    "speed bags",
+    "wool",
+    "control cables",
+    "hand grips",
+    "straight bottom hem",
+    "dumbbell racks",
+    "traction pads",
+    "shuffleboard accessories",
+    "mess kits",
+    "outdoor volleyballs",
+    "goalkeeper equipment",
+    "thailand/vietnam",
+    "thermoses",
+    "elk",
+    "diving packages",
+    "insert tape",
+    "wipe with damp cloth",
+    "80/20 cotton/polyester fleece",
+    "trailer lights",
+    "stainless steel case with stainless-steel band",
+    "portable",
+    "neck gaiters",
+    "meet equipment",
+    "trunks",
+    "floor mats",
+    "trailer winches",
+    "rim brake sets",
+    "paddleboard accessories",
+    "open-hem sleeves; adidas performance",
+    "fish scales",
+    "uniquely designed team slippers",
+    "discuses",
+    "lantern accessories",
+    "progressive fashion with vanguard sports cultures",
+    "reel replacement parts",
+    "dock guards",
+    "fuel-flow instruments",
+    "spokes",
+    "shackles",
+    "t-ball bats",
+    "pack pockets",
+    "made in the usa!",
+    "indoor volleyballs",
+    "expedition tents",
+    "megaphones",
+    "rigid forks",
+    "trainer accessories",
+    "screen printed in the usa",
+    "not water resistant",
+    "nose clips",
+    "full zip hoody",
+    "adult/unisex.",
+    "bocce sets",
+    "marker boards",
+    "skimboards",
+    "freeride boards",
+    "embroidered team logo",
+    "gun rails",
+    "longbows",
+    "sapphire crystal",
+    "poker tables",
+    "screen print logo",
+    "gun racks",
+    "wrist guards",
+    "track jackets",
+    "longeing equipment",
+    "chin straps",
+    "authentic",
+    "valances",
+    "fitness trampolines",
+    "casino prize wheels",
+    "aquatic gloves",
+    "billiards",
+    "unicycles",
+    "stringing kits",
+    "pegs",
+    "set-in sleeves",
+    "made by majestic",
+    "scratch-resistant-sapphire crystal",
+    "quality products",
+    "detailed graphic artwork design",
+    "paddling clothing",
+    "shelters",
+    "laces",
+    "pulleys",
+    "horseshoes",
+    "officially licensed by nfl",
+    "saddle blankets",
+    "lightweight",
+    "chronograph",
+    "bingo sets",
+    "remotes",
+    "machine wash cold",
+    "tire repair kits",
+    "golf bracelets",
+    "waterskis",
+    "mascot costumes",
+    "machine wash cold; tumble dry low",
+    "replacement cleats",
+    "wire terminals",
+    "compression sacks",
+    "card shufflers",
+    "boomerangs",
+    "jumping trainers",
+    "quality embroidery",
+    "trailer accessories",
+    "rib protectors",
+    "full athletic fit",
+    "pellets",
+    "suspension service parts",
+    "parts kits",
+    "officially licensed by the mlb",
+    "adjustable",
+    "boxing pads",
+    "dri-power moisture wicking fabric",
+    "toggle switches",
+    "pogo sticks",
+    "race equipment",
+    "stainless-steel case; black dial; date function",
+    "enclosures",
+    "leg warmers",
+    "licensed collegiate product",
+    "tap dancing equipment",
+    "duffles",
+    "rooftop racks",
+    "water pressure pumps",
+    "japanese quartz movement with analog display",
+    "skorts",
+    "curtains",
+    "in-ground",
+    "card sleeves",
+    "hitting mats",
+    "card albums",
+    "binoculars",
+    "desert dry moisture management polo",
+    "storage",
+    "swim fins",
+    "standard weight benches",
+    "stainless steel case",
+    "attacker shafts",
+    "fitness planners",
+    "scorebooks",
+    "silk eye bags",
+    "wire",
+    "air fresheners",
+    "gun cloths",
+    "dock cleats",
+    "desk lamps",
+    "quartz movement with analog display",
+    "field player gloves",
+    "fast-pitch softballs",
+    "bat racks",
+    "optics covers",
+    "precise and high-quality swiss-quartz movement",
+    "night lights",
+    "quickdraws",
+    "capri pants",
+    "electronic basketball",
+    "foam pads",
+    "personal care products",
+    "mirrors",
+    "machine washable, tumble dry low heat",
+    "regular fit",
+    "spikes",
+    "candles",
+    "facemasks",
+    "pucks",
+    "football helmets",
+    "flag football belts",
+    "treadmill motors",
+    "scent accessories",
+    "handbags",
+    "paperweights",
+    "toasters",
+    "standard bases",
+    "wind socks",
+    "mask defoggers",
+    "caster boards",
+    "finger tabs",
+    "fabric",
+    "spray skirts",
+    "bib shorts",
+    "high quality",
+    "drysuits",
+    "tagless label, eliminates scratching",
+    "double needle stitching",
+    "ball storage",
+    "tank covers",
+    "messenger bags",
+    "precise japanese-quartz movement",
+    "pull-on closure",
+    "spoke decorations",
+    "weapon cases",
+    "billiard lighting",
+    "power cages",
+    "sliding pads",
+    "short sleeve basic crew neck tee",
+    "buckle closure",
+    "slap straps",
+    "risers",
+    "raglan sleeves",
+    "support your favorite nhl team",
+    "parts",
+    "ncaa licensed product",
+    "tie-downs",
+    "fitness circles",
+    "catcher helmets",
+    "table tennis sets",
+    "water sports accessories",
+    "gaffs",
+    "jacks",
+    "exercise ball accessories",
+    "forearm guards",
+    "adjustable plastic snapback cap",
+    "swim belts",
+    "face guards",
+    "poker table tops",
+    "padded shorts",
+    "officially licensed product",
+    "snorkel vests",
+    "hearing enhancement aids",
+    "unstrung",
+    "chain guards",
+    "made by new era",
+    "cotton blend",
+    "polyester-blend",
+    "rear shocks",
+    "by adidas",
+    "quickly",
+    "tetherball",
+    "gun snakes",
+    "goose",
+    "net systems",
+    "climbing protection",
+    "shot puts",
+    "training cones",
+    "foam wedges",
+    "marker buoys",
+    "windsurfing",
+    "agility ladders",
+    "training putters",
+    "polycarbonate lens",
+    "kneeboarding",
+    "cornhole boards",
+    "clips",
+    "short sleeves",
+    "desert dry polo",
+    "basic bows",
+    "fleece lining",
+    "tip-ups",
+    "gun holders",
+    "outdoor skates",
+    "short-sleeve shirt",
+    "fishing belts",
+    "co2 systems",
+    "slim flights",
+    "dri-fit",
+    "sweatpants",
+    "indoor surfboard storage",
+    "water helmets",
+    "elastic waistband",
+    "dip stands",
+    "calorie counters",
+    "horse driving equipment",
+    "ladder ball",
+    "performance wear, loose fit",
+    "swiss quartz movement with analog display",
+    "knee high socks",
+    "digital gauges",
+    "indoor kayak storage",
+    "shoulder-to-shoulder taping",
+    "adjustable closure",
+    "wall-mount",
+    "eggs",
+    "roulette equipment",
+    "dog packs",
+    "olympic weight benches",
+    "surfboard racks",
+    "pet harnesses",
+    "machine wash, tumble dry",
+    "made in pakistan",
+    "hecho en ee. uu. e importado",
+    "side entry pockets",
+    "child seats",
+    "officially licensed.",
+    "computer accessories",
+    "self fabric neck band",
+    "utility cord",
+    "spoke tools",
+    "durable mineral crystal",
+    "nipples",
+    "crops",
+    "made with techno-dry",
+    "ankle guards",
+    "shortboards",
+    "women's balls",
+    "classic style hood with drawstring",
+    "official nba product",
+    "tagless heat-seal neck label",
+    "hockey sticks",
+    "padding supplies",
+    "grease",
+    "weather monitors",
+    "external frame backpacks",
+    "great gift for sport fans!",
+    "automatic movement",
+    "pockets",
+    "coverseamed neck double-needle hem",
+    "cotton-blend",
+    "spike wrenches",
+    "unisex",
+    "composite frame",
+    "kicking targets",
+    "crewneck",
+    "bells",
+    "replacement lenses",
+    "preshrunk",
+    "premium knit cap",
+    "39thirty flex fit cap",
+    "mitt accessories",
+    "water globes",
+    "workout shorts",
+    "analog-quartz movement",
+    "heatgear",
+    "machine wash and dry",
+    "sleeveless suits",
+    "cotton/polyester",
+    "pouch pocket",
+    "scoop neck tee",
+    "rib knit cuffs and waistband",
+    "golf gift sets",
+    "motor flushers",
+    "patented one-fit sizing",
+    "tagless technology for a better fit",
+    "nike",
+    "strung",
+    "magnified date window at 3:00",
+    "outdoor thermometers",
+    "sais",
+    "double-end bags",
+    "fly tying kits",
+    "plastic balls",
+    "bean bag chairs",
+    "moisture-wicking finish",
+    "kicking shields",
+    "embroidered logo",
+    "two front pockets",
+    "reversible",
+    "thailand - vietnam",
+    "bangladesh/pakistan",
+    "material:spandex",
+    "goal targets",
+    "canoes",
+    "steering wheels",
+    "india",
+    "waterproof",
+    "comfortable",
+    "mailboxes",
+    "black",
+    "fixed gear bike frames",
+    "coverstiched seams",
+    "canvas",
+    "paper cups",
+    "snap back",
+    "no fur",
+    "great quality jewelry!",
+    "perfect for every day use",
+    "nylon lining",
+    "chain tension idlers",
+    "alarm",
+    "machine wash cold, air dry",
+    "mounts",
+    "rails",
+    "japanese-quartz movement",
+    "lesotho",
+    "embroidered team graphics"
+]

interactive/pyvis_graph.py

@@ -117,7 +117,8 @@ def generate_network(kb, node_id, max_nodes=10, num_hops='2'):
     if num_hops == "2":
         edge_index, edge_types = get_one_hop(kb, node_id, max_nodes)
         neighbor_nodes = torch.unique(edge_index).tolist()
-        neighbor_nodes.remove(node_id)
+        if node_id in neighbor_nodes:
+            neighbor_nodes.remove(node_id)
 
         for neighbor_node in neighbor_nodes:
             e_index, e_type = get_one_hop(kb, neighbor_node, max_nodes=1)
@@ -134,6 +135,8 @@ def generate_network(kb, node_id, max_nodes=10, num_hops='2'):
         edge_index = torch.cat([edge_index, add_edge_index], dim=1)
         edge_types = torch.cat([edge_types, add_edge_types], dim=0)
 
+    # add a self-loop for node_id to avoid isolated node
+    edge_index = torch.concat([edge_index, torch.LongTensor([[node_id], [node_id]])], dim=1)
     node_ids, relabel_edge_index, _, _ = relabel(
         torch.arange(kb.num_nodes()), edge_index, batch=torch.zeros(kb.num_nodes())
     )
@@ -157,6 +160,8 @@ def generate_network(kb, node_id, max_nodes=10, num_hops='2'):
                 font={"align": "middle", "size": 10},
             )
     for idx in range(relabel_edge_index.size(-1)):
+        if relabel_edge_index[0][idx].item() == relabel_edge_index[1][idx].item():
+            continue
         if indirected:
             net.add_edge(
                 relabel_edge_index[0][idx].item(),
@@ -195,7 +200,7 @@ def get_text_html(kb, node_id):
         font-family: SF Pro Text, SF Pro Icons, Helvetica Neue, Helvetica, Arial, sans-serif;">{text}</div>"""
 
 
-def get_subgraph_html(kb, kb_name, node_id, max_nodes=10, num_hops=1):
+def get_subgraph_html(kb, kb_name, node_id, max_nodes=10, num_hops='1'):
     network = generate_network(kb, node_id, max_nodes, num_hops)
 
     nodes = network[0]
@@ -271,5 +276,4 @@ def main():
 
 
 if __name__ == "__main__":
-
     main()

src/benchmarks/get_semistruct.py

@@ -7,17 +7,18 @@ def get_semistructured_data(name, root='data/', download_processed=True, **kwarg
     if name == 'amazon':
         categories = ['Sports_and_Outdoors']
         kb = AmazonSemiStruct(root=data_root,
-                                categories=categories,
-                                meta_link_types=['brand'],
-                                download_processed=download_processed,
-                                **kwargs
-                                )
+                              categories=categories,
+                              meta_link_types=['brand', 'category'],
+                              download_processed=download_processed,
+                              **kwargs
+                              )
     if name == 'primekg':
         kb = PrimeKGSemiStruct(root=data_root, 
-                               download_processed=download_processed,
-                               **kwargs)
+                                download_processed=download_processed,
+                                **kwargs)
     
     if name == 'mag':
         kb = MagSemiStruct(root=data_root, 
-                           download_processed=download_processed)
-    return kb
+                            download_processed=download_processed,
+                            **kwargs)
+    return kb

src/benchmarks/semistruct/amazon.py

@@ -20,7 +20,6 @@ PROCESSED_DATASET = {
     "file": "amazon_processed.zip",
 }
 
-
 class AmazonSemiStruct(SemiStructureKB):
     
     REVIEW_CATEGORIES = set(['Amazon_Fashion','All_Beauty','Appliances',
@@ -49,6 +48,8 @@ class AmazonSemiStruct(SemiStructureKB):
                   'Office_Products', 'Patio_Lawn_and_Garden', 'Pet_Supplies', 'Sports_and_Outdoors', 
                   'Tools_and_Home_Improvement', 'Toys_and_Games', 'Video_Games'])
     
+    sub_category = 'data/amazon/stats/category_list.json'
+    SUB_CATEGORIES = set(json.load(open(sub_category, 'r')))
     link_columns = ['also_buy', 'also_view']
     review_columns = ['reviewerID', 'summary', 'reviewText', 'vote', 'overall', 'verified', 'reviewTime']
     qa_columns = ['questionType', 'answerType', 'question', 'answer', 'answerTime']
@@ -56,12 +57,13 @@ class AmazonSemiStruct(SemiStructureKB):
                     'rank', 'details', 'description']
     candidate_types = ['product']
     node_attr_dict = {'product': ['title', 'dimensions', 'weight', 'description', 'features', 'reviews', 'Q&A'],
-                       'brand': ['brand_name']}
+                       'brand': ['brand_name'],
+                       'category': ['category_name']}
 
     def __init__(self, 
                  root,
                  categories: list, 
-                 meta_link_types=['brand'],
+                 meta_link_types=['category'],
                  max_entries=25,
                  download_processed=True,
                  **kwargs):
@@ -104,6 +106,8 @@ class AmazonSemiStruct(SemiStructureKB):
             print(f'Load cached graph with meta link types {meta_link_types}')
             processed_data = load_files(cache_path)
         else:
+            print(f'Start processing raw data...')
+            print(f'{meta_link_types=}')
             processed_data = self._process_raw(categories)
             if meta_link_types: 
                 # customize the graph by adding meta links
@@ -113,10 +117,10 @@ class AmazonSemiStruct(SemiStructureKB):
     def __getitem__(self, idx):
         idx = int(idx)
         node_info = self.node_info[idx]
-        try:
-            dimensions, weight = node.details.dictionary.product_dimensions.split(' ; ')
-            node_info['dimensions'], node_info['weight'] = dimensions, weight
-        except: pass
+        # try:
+        #     dimensions, weight = node.details.dictionary.product_dimensions.split(' ; ')
+        #     node_info['dimensions'], node_info['weight'] = dimensions, weight
+        # except: pass
         node = Node()
         register_node(node, node_info)
         return node
@@ -167,6 +171,8 @@ class AmazonSemiStruct(SemiStructureKB):
         
         if self.node_type_dict[int(self.node_types[idx])] == 'brand':
             return f'brand name: {self[idx].brand_name}'
+        if self.node_type_dict[int(self.node_types[idx])] == 'category':
+            return f'category name: {self[idx].category_name}'
         
         node = self[idx]
         doc = f'- product: {node.title}\n'
@@ -268,29 +274,50 @@ class AmazonSemiStruct(SemiStructureKB):
         print(f'Check data downloading...')
         for category in review_categories:
             review_header = 'https://datarepo.eng.ucsd.edu/mcauley_group/data/amazon_v2'
-            download_url(f'{review_header}/categoryFiles/{category}.json.gz', self.raw_data_dir)
-            download_url(f'{review_header}/metaFiles2/meta_{category}.json.gz', self.raw_data_dir)
+            if not os.path.exists(osp.join(self.raw_data_dir, f'{category}.json.gz')):
+                print(f'Downloading {category} data...')
+                download_url(f'{review_header}/categoryFiles/{category}.json.gz', self.raw_data_dir)
+                download_url(f'{review_header}/metaFiles2/meta_{category}.json.gz', self.raw_data_dir)
         for category in qa_categories:
             qa_header = 'https://datarepo.eng.ucsd.edu/mcauley_group/data/amazon/qa'
-            download_url(f'{qa_header}/qa_{category}.json.gz', self.raw_data_dir)
+            if not os.path.exists(osp.join(self.raw_data_dir, f'qa_{category}.json.gz')):
+                print(f'Downloading {category} QA data...')
+                download_url(f'{qa_header}/qa_{category}.json.gz', self.raw_data_dir)
             
         if not osp.exists(osp.join(self.processed_data_dir, 'node_info.pkl')):
+            ckt_path = 'data/amazon/intermediate'
             print('Loading data... It might take a while')
             # read amazon QA data
-            df_qa = pd.concat([read_qa(osp.join(self.raw_data_dir, f'qa_{category}.json.gz'))
-                               for category in qa_categories])[['asin'] + self.qa_columns]
-            
+            df_qa_path = os.path.join(ckt_path, 'df_qa.pkl')
+            if os.path.exists(df_qa_path):
+                df_qa = pd.read_pickle(df_qa_path)
+            else:
+                df_qa = pd.concat([read_qa(osp.join(self.raw_data_dir, f'qa_{category}.json.gz'))
+                                for category in qa_categories])[['asin'] + self.qa_columns]
+                df_qa.to_pickle(df_qa_path)
+            print('df_qa loaded')
             # read amazon review data
-            df_review = pd.concat([read_review(osp.join(self.raw_data_dir, f'{category}.json.gz')) 
+            df_review_path = os.path.join(ckt_path, 'df_review.pkl')
+            if os.path.exists(df_review_path):
+                df_review = pd.read_pickle(df_review_path)
+            else:
+                df_review = pd.concat([read_review(osp.join(self.raw_data_dir, f'{category}.json.gz')) 
                                    for category in review_categories])[['asin'] + self.review_columns]
+                df_review.to_pickle(df_review_path)
+            print('df_review loaded')
             # read amazon meta data from amazon review & amazon kdd
-            meta_df_lst = []
-            for category in review_categories:
-                cat_review = read_review(osp.join(self.raw_data_dir, f'meta_{category}.json.gz'))
-                cat_review.insert(0, 'global_category', category.replace('_', ' '))
-                meta_df_lst.append(cat_review)
-            df_ucsd_meta = pd.concat(meta_df_lst)
-            
+            df_ucsd_meta_path = os.path.join(ckt_path, 'df_ucsd_meta.pkl')
+            if os.path.exists(df_ucsd_meta_path):
+                df_ucsd_meta = pd.read_pickle(df_ucsd_meta_path)
+            else:
+                meta_df_lst = []
+                for category in review_categories:
+                    cat_review = read_review(osp.join(self.raw_data_dir, f'meta_{category}.json.gz'))
+                    cat_review.insert(0, 'global_category', category.replace('_', ' '))
+                    meta_df_lst.append(cat_review)
+                df_ucsd_meta = pd.concat(meta_df_lst)
+                df_ucsd_meta.to_pickle(df_ucsd_meta_path)
+            print('df_ucsd_meta loaded')
             print('Preprocessing data...')
             df_ucsd_meta = df_ucsd_meta.drop_duplicates(subset='asin', keep='first')
             df_meta = df_ucsd_meta[self.meta_columns + self.link_columns]
@@ -342,8 +369,23 @@ class AmazonSemiStruct(SemiStructureKB):
         
         n_e_types, n_n_types = len(edge_type_dict), len(node_type_dict)
         for i, link_type in enumerate(meta_link_types):
-            values = np.array([self._process_brand(node_info_i[link_type]) for node_info_i in node_info.values() if link_type in node_info_i.keys()])
-            indices = np.array([idx for idx, node_info_i in enumerate(node_info.values()) if link_type in node_info_i.keys()])
+            if link_type == 'brand':
+                values = np.array([self._process_brand(node_info_i[link_type]) for node_info_i in node_info.values() if link_type in node_info_i.keys()])
+                indices = np.array([idx for idx, node_info_i in enumerate(node_info.values()) if link_type in node_info_i.keys()])
+            elif link_type == 'category':
+                value_list = []
+                indice_list = []
+                for idx, node_info_i in enumerate(node_info.values()):
+                    if link_type in node_info_i.keys():
+                        value_list.extend(node_info_i[link_type])
+                        indice_list.extend([idx for _ in range(len(node_info_i[link_type]))])
+                values = np.array(value_list)
+                indices = np.array(indice_list)
+                print(f'{link_type=}, {len(values)=}, {len(indices)=}')
+                # print(values[:50])
+                print(indices[:50])
+            else:
+                raise Exception(f'Invalid meta link type {link_type}')
             
             cur_n_nodes = len(node_info)
             node_type_dict[n_n_types + i] = link_type
@@ -398,6 +440,14 @@ class AmazonSemiStruct(SemiStructureKB):
                     brand = self._process_brand(clean_data(df_meta_i[column]))
                     if len(brand) > 1:
                         node_info[idx]['brand'] = brand
+                elif column == 'category':
+                    category_list = []
+                    for category in df_meta_i[column]:
+                        category = category.lower()
+                        if category in self.SUB_CATEGORIES:
+                            category_list.append(category)
+                    if len(category_list) > 0:
+                        node_info[idx]['category'] = category_list
                 else:
                     node_info[idx][column] = clean_data(df_meta_i[column])
                         
@@ -482,4 +532,4 @@ def read_qa(path):
       df[i] = d
       i += 1
     return pd.DataFrame.from_dict(df, orient='index')
-  return getDF(path)
+  return getDF(path)