wehub-resource-sync
c3749daf48
Tests / test-linux (3.13) (push) Failing after 0s
Tests / test-linux (3.11) (push) Failing after 1s
Tests / lint (push) Failing after 0s
Tests / test-linux (3.9) (push) Failing after 1s
Docker / build (push) Failing after 1s
Docker / build-gpu (push) Failing after 2s
Tests / test-windows (push) Has been cancelled
Tests / test-macos (push) Has been cancelled
134 lines
11 KiB
JSON
134 lines
11 KiB
JSON
{
|
|
"schema_version": 1,
|
|
"description": "Top common English content words filtered from entity candidacy. These are nouns, adjectives, and verbs that frequently appear capitalized (at sentence-start, in headings, in markdown) but are NOT proper nouns. Without this filter they get false-positive detected as people or projects. POS coverage: common noun (n), adjective (adj), verb (v). Excludes: proper nouns, known product/technology names (those are protected separately by the planned known-systems lexicon).",
|
|
"source": "Curated from (1) observed false positives in real mempalace palaces, (2) COCA top-2000 content-word frequency list (POS-filtered to noun/adj/verb).",
|
|
"case_handling": "Matching is case-insensitive. Words stored lowercase. Caller must lowercase candidate before lookup.",
|
|
"tier_3_interaction": "When the known-systems lexicon ships (Tier 3), candidates that appear as part of a multi-word compound in that lexicon should be allowed even if their single-word form is in this list. e.g., 'Code' alone is filtered; 'Claude Code' as a phrase is preserved.",
|
|
"words": [
|
|
"able", "about", "above", "across", "act", "action", "actually", "add", "address",
|
|
"after", "again", "age", "agency", "ago", "agree", "ahead", "air", "all", "allow",
|
|
"almost", "alone", "along", "already", "also", "although", "always", "among",
|
|
"amount", "another", "answer", "anti", "any", "anything", "anyway", "appear",
|
|
"apply", "approach", "area", "around", "art", "article", "ask", "audit",
|
|
"available", "average", "away", "back", "background", "backup", "backups", "bad",
|
|
"base", "basic", "be", "because", "become", "before", "begin", "behind",
|
|
"believe", "below", "best", "better", "between", "big", "bigger", "black", "block",
|
|
"blocks", "blue", "body", "book", "both", "bottom", "boy", "branch", "break",
|
|
"bring", "broad", "brutal", "brutally", "budget", "build", "building", "built",
|
|
"business", "but", "buy", "call", "came", "can", "cannot", "card", "care",
|
|
"career", "case", "cases", "category", "cause", "cell", "center", "central",
|
|
"century", "certain", "certainly", "chain", "chance", "change", "changes",
|
|
"channel", "chat", "check", "child", "children", "choice", "choose", "claim",
|
|
"class", "clean", "clear", "clearly", "click", "client", "close", "code",
|
|
"coder", "coding", "color", "column", "combine", "come", "command", "comment",
|
|
"common", "community", "company", "compare", "complete", "concept", "concern",
|
|
"concerns", "condition", "configure", "confirm", "consider", "contain",
|
|
"container", "content", "context", "continue", "contract", "control", "convert",
|
|
"core", "corner", "correct", "cost", "could", "count", "country", "couple",
|
|
"course", "cover", "create", "current", "currently", "cut", "data", "date",
|
|
"day", "days", "deal", "debug", "decade", "decide", "decision", "deep",
|
|
"default", "define", "delete", "deliver", "demand", "deny", "depend", "deploy",
|
|
"describe", "design", "detail", "details", "develop", "did", "die", "difference",
|
|
"different", "difficult", "direct", "directory", "discover", "discuss",
|
|
"discussion", "do", "doc", "docs", "does", "doing", "domain", "done", "door",
|
|
"down", "draft", "drawer", "drop", "due", "during", "each", "early", "easy",
|
|
"edge", "edit", "education", "effect", "effort", "either", "else", "email",
|
|
"empty", "end", "energy", "engine", "english", "enjoy", "enough", "enter",
|
|
"entire", "environment", "error", "errors", "establish", "even", "evening",
|
|
"event", "every", "everyone", "everything", "exactly", "example", "examples",
|
|
"except", "exist", "expand", "expect", "experience", "explain", "extract",
|
|
"eye", "face", "fact", "facts", "fail", "failure", "fall", "false", "family",
|
|
"fast", "father", "feature", "feed", "feel", "field", "fight", "figure", "file",
|
|
"files", "fill", "final", "finally", "find", "fine", "finish", "first", "fix",
|
|
"fixed", "flag", "flags", "flat", "flow", "focus", "folder", "follow", "food",
|
|
"foot", "for", "force", "form", "format", "forward", "found", "four", "free",
|
|
"friend", "from", "front", "full", "function", "future", "game", "general",
|
|
"generate", "get", "girl", "give", "given", "global", "go", "goal", "going",
|
|
"gone", "good", "got", "government", "great", "green", "ground", "group",
|
|
"grow", "growth", "guess", "guide", "had", "half", "hand", "handle", "happen",
|
|
"happy", "hard", "has", "have", "head", "header", "health", "hear", "heart",
|
|
"heavy", "held", "help", "her", "here", "high", "him", "his", "history", "hit",
|
|
"hold", "home", "hook", "hope", "hot", "hour", "house", "how", "however",
|
|
"human", "huge", "hundred", "idea", "identify", "if", "image", "imagine",
|
|
"impact", "implement", "important", "improve", "include", "increase", "index",
|
|
"indicate", "industry", "info", "information", "init", "input", "inside",
|
|
"install", "instead", "interest", "internal", "introduce", "investment",
|
|
"investigate", "involve", "issue", "issues", "it", "item", "items", "its",
|
|
"itself", "job", "join", "just", "keep", "kept", "key", "kid", "kill", "kind",
|
|
"knew", "know", "knowledge", "known", "label", "language", "large", "last",
|
|
"late", "later", "law", "lawyer", "lay", "lead", "leader", "learn", "least",
|
|
"leave", "led", "left", "legal", "less", "let", "letter", "level", "lie",
|
|
"life", "light", "like", "likely", "line", "link", "list", "listen", "little",
|
|
"live", "load", "local", "log", "logs", "long", "longer", "look", "loop",
|
|
"lose", "loss", "lost", "lot", "love", "low", "machine", "made", "main",
|
|
"major", "make", "man", "manage", "many", "map", "mar", "market", "match",
|
|
"matter", "may", "maybe", "me", "mean", "meant", "media", "meet", "meeting",
|
|
"member", "memory", "merge", "message", "method", "middle", "might", "million",
|
|
"mind", "minute", "mission", "model", "modify", "moment", "money", "month",
|
|
"more", "morning", "most", "mother", "move", "movement", "much", "music",
|
|
"must", "my", "myself", "name", "names", "national", "natural", "near", "need",
|
|
"needed", "needs", "neither", "network", "never", "new", "news", "next",
|
|
"nice", "night", "no", "none", "nor", "north", "not", "note", "notes",
|
|
"nothing", "notice", "now", "number", "object", "obvious", "obviously", "of",
|
|
"off", "offer", "office", "official", "often", "oh", "ok", "okay", "old",
|
|
"on", "once", "one", "only", "open", "operation", "opinion", "opportunity",
|
|
"option", "options", "or", "order", "organization", "original", "other",
|
|
"others", "our", "out", "outside", "over", "own", "page", "pages", "pair",
|
|
"paper", "parent", "part", "particular", "particularly", "partner", "party",
|
|
"pass", "passed", "past", "patch", "path", "pattern", "pay", "peace", "people",
|
|
"per", "performance", "perhaps", "period", "person", "personal", "phase",
|
|
"phases", "phone", "physical", "pick", "piece", "place", "plan", "planning",
|
|
"play", "please", "point", "police", "policy", "political", "poor", "popular",
|
|
"population", "port", "position", "positive", "possible", "post", "power",
|
|
"pre", "prepare", "present", "president", "press", "pretty", "prevent",
|
|
"previous", "previously", "price", "primary", "print", "private", "probably",
|
|
"problem", "problems", "process", "produce", "product", "professional",
|
|
"program", "project", "promise", "property", "protect", "prove", "provide",
|
|
"public", "pull", "purpose", "push", "put", "question", "quick", "quickly",
|
|
"quite", "quote", "race", "raise", "range", "rather", "reach", "read", "ready",
|
|
"real", "reality", "realize", "really", "reason", "receive", "recent",
|
|
"recently", "recognize", "record", "red", "reduce", "ref", "reference",
|
|
"refresh", "region", "release", "remain", "remember", "remove", "render",
|
|
"repeat", "replace", "reply", "report", "represent", "request", "require",
|
|
"research", "resource", "respond", "response", "rest", "result", "results",
|
|
"return", "review", "rich", "right", "rise", "risk", "road", "role", "room",
|
|
"rooms", "rule", "rules", "run", "running", "safe", "said", "same", "save",
|
|
"saw", "say", "scale", "scene", "school", "science", "scope", "score", "screen",
|
|
"scroll", "sea", "search", "season", "second", "section", "see", "seek", "seem",
|
|
"seen", "select", "self", "sell", "send", "sense", "sent", "serious", "serve",
|
|
"service", "session", "set", "settle", "seven", "several", "shake", "shape",
|
|
"share", "she", "ship", "shoot", "short", "should", "show", "side", "sign",
|
|
"signal", "significant", "similar", "simple", "simply", "since", "sing",
|
|
"single", "sister", "sit", "site", "situation", "six", "size", "skip", "skill",
|
|
"skin", "sky", "sleep", "slip", "slot", "slow", "small", "smile", "snippet",
|
|
"so", "social", "society", "soft", "software", "soldier", "solid", "solution",
|
|
"solve", "some", "somebody", "somehow", "someone", "something", "sometimes",
|
|
"somewhere", "son", "song", "soon", "sort", "sound", "source", "space", "speak",
|
|
"special", "specific", "spend", "spent", "split", "sport", "spread", "stable",
|
|
"stack", "stage", "stand", "standard", "star", "start", "state", "statement",
|
|
"station", "stay", "step", "still", "stock", "stop", "store", "story", "stream",
|
|
"street", "strict", "strong", "structure", "student", "study", "stuff",
|
|
"subject", "success", "such", "sudden", "suggest", "summary", "support",
|
|
"suppose", "sure", "surface", "switch", "system", "table", "take", "talk",
|
|
"tall", "target", "task", "tax", "teach", "team", "tell", "ten", "tend",
|
|
"term", "test", "tests", "text", "than", "thank", "that", "the", "their",
|
|
"them", "themselves", "then", "theory", "there", "these", "they", "thing",
|
|
"things", "think", "third", "this", "those", "though", "thought", "thousand",
|
|
"thread", "three", "through", "throw", "thus", "tier", "ticket", "time",
|
|
"title", "to", "today", "together", "too", "tool", "top", "topic", "total",
|
|
"touch", "toward", "tower", "town", "track", "trade", "train", "transfer",
|
|
"treat", "tree", "trial", "trigger", "trip", "trouble", "true", "truly",
|
|
"trust", "truth", "try", "turn", "tweak", "two", "type", "under", "understand",
|
|
"unit", "until", "up", "update", "upload", "upon", "us", "usage", "use", "used",
|
|
"user", "usually", "valid", "value", "various", "verb", "verbose", "verify",
|
|
"version", "very", "via", "view", "visit", "voice", "vote", "wait", "walk",
|
|
"wall", "want", "war", "warn", "watch", "water", "way", "we", "weak", "wear",
|
|
"web", "week", "weight", "well", "west", "what", "whatever", "when", "where",
|
|
"whether", "which", "while", "white", "who", "whole", "whose", "why", "wide",
|
|
"wife", "will", "win", "wind", "window", "wing", "wire", "wish", "with",
|
|
"within", "without", "woman", "word", "work", "worker", "world", "worry",
|
|
"would", "write", "writer", "wrong", "yard", "yeah", "year", "yes", "yet",
|
|
"you", "young", "your", "yours", "yourself", "zero", "zone"
|
|
]
|
|
}
|