Class: OKF::CLI::Search
Overview
Ranked text retrieval — the browser page's search brought to the CLI on the same engine (a MiniFTS index) and extended to bodies. Terms after the directory are ANDed tokens, matched whole or by prefix (Ruby regexps with --regexp, typo tolerance with --fuzzy); rows rank by BM25+ weighted toward where they hit (title > id > tags > type/description > body) and carry one bounded context snippet, so "which concept covers X?" costs a few rows, not a body read. Advisory read: exit 0 even with no matches. Exact by default — the consuming agent is the fuzzy layer, until it asks not to be.
Constant Summary collapse
- CAPABILITY_FLAGS =
The core raises
:regexp; a user typed--regexp. Translating here keeps the flag vocabulary in the shell, where it belongs, and lets the message end with the fix rather than only the complaint: an engine that can do what was asked is named, so the next command is obvious. { regexp: "--regexp", fuzzy: "--fuzzy" }.freeze
Constants inherited from Command
Command::DUCK_TYPE, Command::FILTER_KEYS
Class Method Summary collapse
Instance Method Summary collapse
Methods inherited from Command
Constructor Details
This class inherits a constructor from OKF::CLI::Command
Class Method Details
.group ⇒ Object
24 25 26 |
# File 'lib/okf/cli/search.rb', line 24 def self.group :read end |
.help_rows ⇒ Object
28 29 30 31 32 |
# File 'lib/okf/cli/search.rb', line 28 def self.help_rows [ [ "search <dir|@slug…|@all> <term…> [--regexp|--fuzzy]", "find concepts by text or regexp, ranked (@all: every bundle)" ] ] end |
.id ⇒ Object
20 21 22 |
# File 'lib/okf/cli/search.rb', line 20 def self.id :search end |
Instance Method Details
#call(argv) ⇒ Object
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 |
# File 'lib/okf/cli/search.rb', line 34 def call(argv) = { json: false, regexp: false, fuzzy: false, engine: nil } parser = OptionParser.new do |o| o. = "Usage: okf search <dir|@slug…|@all> <term…> [--engine NAME] [--regexp|--fuzzy] " \ "[--in FIELDS] [--type T] [--dir D] [--tag T] [--status S] [--trust T] [--json]" search_engine_note(o) json_flags(o, , "emit the matches as JSON") projection_flags(o, ) o.on("-e", "--regexp", "read each term as a Ruby regular expression rather", "than literal text — case-insensitive (scan engine)") { [:regexp] = true } o.on("--fuzzy", "tolerate typos, edit distance #{OKF::Bundle::Search::FUZZY_DISTANCE} × term length (index engine)") { [:fuzzy] = true } o.on("--engine NAME", "match with this engine instead of the default", "(#{engine_names}) — index is BM25+ ranked, token-based") { |v| [:engine] = v } o.on("--in LIST", Array, "search only these fields (#{OKF::Bundle::Search::FIELDS.join(", ")})") { |v| [:in] = v.map(&:downcase) } filter_flags(o, , :type, :area, :tag) help_flag(o) end begin parser.parse!(argv) rescue OptionParser::ParseError => e @err.puts e. return 2 end # Registry mode — leading @refs, @all among them — searches several bundles # and labels every match; a plain dir keeps the classic single-bundle output. if argv.first&.start_with?("@") pairs = ref_targets(argv) or return 2 dir = nil else dir = argv.shift if dir.nil? @err.puts parser. return 2 end dir = resolve_ref(dir) or return 2 end terms = argv if terms.empty? @err.puts parser. return 2 end # A non-leading @arg is a literal term by the grammar — say so, since the # user may have meant a ref (refs must lead) and would otherwise see only # a silent zero-match. stray = terms.find { |term| term.start_with?("@") } @err.puts "note: '#{stray}' searches as a literal term — an @slug or @all must lead" if stray unknown = Array([:in]) - OKF::Bundle::Search::FIELDS return usage_error("unknown field(s): #{unknown.join(", ")} (searchable: #{OKF::Bundle::Search::FIELDS.join(", ")})") unless unknown.empty? # Two query languages, not two dials on one: a regexp is matched against raw # text, --fuzzy is an edit distance over indexed tokens. Silently honouring # one and dropping the other would answer a question nobody asked. if [:regexp] && [:fuzzy] return usage_error("--regexp and --fuzzy are mutually exclusive (a pattern is matched literally, not by edit distance)") end return multi_search(pairs, terms, ) if pairs folder = OKF::Bundle::Folder.load(dir) report_skipped(folder) rows = OKF::Bundle::Search.call(folder.bundle, terms, fields: [:in], regexp: [:regexp], fuzzy: [:fuzzy], engine: [:engine]) keep = filter_ids(folder, ) rows = rows.select { |row| keep.include?(row[:id]) } unless keep.nil? return print_search_json(dir, terms, rows, ) if [:json] print_search(dir, terms, rows, folder.bundle.concepts.size) 0 rescue RegexpError => e usage_error("invalid pattern: #{e.}") rescue OKF::Bundle::Search::UnknownEngine => e usage_error(e.) rescue OKF::Bundle::Search::UnsupportedQuery => e usage_error((e)) end |