com.blockether.fff — com.blockether/fff 0.12.0

Clojure binding to fff — fast typo-tolerant file/content search — through fff-c's C ABI using the JDK Foreign Function & Memory API.

base-path clj

(base-path fff)

create clj

(create)

(create {:keys [base-path frecency-db-path history-db-path enable-mmap-cache?
                enable-content-indexing? watch? ai-mode? log-file-path log-level
                cache-budget-max-files cache-budget-max-bytes
                cache-budget-max-file-size enable-fs-root-scanning?
                enable-home-dir-scanning? follow-symlinks? respect-ignore-files?
                custom-ignore-filenames exclude-globs unignore-globs]
         :or {base-path (System/getProperty "user.dir")
              enable-mmap-cache? false
              enable-content-indexing? true
              watch? false
              ai-mode? false
              respect-ignore-files? true}})

Create and return an Fff instance. Options: :base-path (default current dir), :frecency-db-path, :history-db-path, :enable-mmap-cache?, :enable-content-indexing?, :watch?, :ai-mode?, :log-file-path, :log-level, cache budget keys, :follow-symlinks?, root/home scanning flags, and :respect-ignore-files? (default true — set false to index files excluded by .gitignore/.ignore/.git/info/exclude; .git internals stay excluded either way).

Ignore overlay (each a seq of strings, all optional):

• :custom-ignore-filenames — extra per-directory ignore-file names to honor, e.g. [".rgignore"]. The ignore crate registers .gitignore/.ignore itself; anything else must be opted in here.
• :exclude-globs — gitignore-syntax patterns ALWAYS excluded, whatever the ignore files say and even when :respect-ignore-files? is false.
• :unignore-globs — gitignore-syntax patterns RE-INCLUDED although an ignore file excludes them (a whitelist; :exclude-globs still wins). No-op when :respect-ignore-files? is false. The overlay is honored by the scan walk AND by the live watcher, so a re-included subtree stays in the index.

destroy! clj

(destroy! fff)

glob clj

(glob fff
      {:keys [pattern current-file max-threads page-index page-size]
       :or {max-threads 0 page-index 0 page-size 100}})

Glob-only path search. Options: :pattern, :current-file, :max-threads, :page-index, :page-size.

grep clj

(grep fff
      {:keys [query mode max-file-size max-matches-per-file smart-case?
              file-offset page-limit time-budget-ms before-context after-context
              classify-definitions?]
       :or {classify-definitions? false
            page-limit 50
            max-file-size 0
            smart-case? true
            after-context 0
            mode :plain
            before-context 0
            max-matches-per-file 0
            time-budget-ms 0
            file-offset 0}})

Content search. :mode is :plain (default), :regex, or :fuzzy. Other options mirror fff-c: max file size, matches/page limits, smart case, context, definition classification, file pagination, and time budget.

health-check clj

(health-check fff)

(health-check fff test-path)

Return fff-c's health-check JSON string. Pass nil or omit test-path to check the instance base path.

historical-query clj

(historical-query fff offset)

refresh-git-status! clj

(refresh-git-status! fff)

rescan! clj

(rescan! fff)

(rescan! fff timeout-ms)

Rescan the filesystem and BLOCK until the fresh index is queryable.

Read-your-writes: after this returns true, search/grep see files and file contents written before the call. Returns false on timeout.

scan-files! clj

(scan-files! fff)

Trigger an ASYNC rescan. Fire-and-forget: the request is dropped into a pending flag while a previous scan's post-scan indexing is still running, and wait-for-scan returns before the grep content index is rebuilt. Use rescan! when the caller must see its own writes.

search clj

(search fff
        {:keys [query current-file max-threads page-index page-size
                combo-boost-multiplier min-combo-count]
         :or {query ""
              max-threads 0
              page-index 0
              page-size 100
              combo-boost-multiplier 100
              min-combo-count 3}})

Fuzzy path search. Options: :query, :current-file, :max-threads, :page-index, :page-size, :combo-boost-multiplier, :min-combo-count.

track-query! clj

(track-query! fff query file-path)

wait-for-scan clj

(wait-for-scan fff)

(wait-for-scan fff timeout-ms)

Wait until the initial scan completes. Returns true on completion, false on timeout.

with-instance clj macro

(with-instance [sym opts] & body)