Compare commits

...

4 commits

Author SHA1 Message Date
1890f14f9e Implement support for multiple filters
Allows for adding more filters so regex or searching by thread number
will be much easier
2024-09-24 00:58:53 +02:00
18cc3e730c Refactored that horrible abomination of a code
I don't do drugs but I must have or something, otherwise that is just
unexplainable, I'm sorry if you had to see that, I really am
2024-09-24 00:53:47 +02:00
8d61968dc9 Make filters take the whole thread to be more flexible 2024-09-24 00:28:04 +02:00
817790cfb4 Fix repl-main, add bunch of TODOs 2024-09-24 00:26:28 +02:00
5 changed files with 40 additions and 41 deletions

View file

@ -78,6 +78,8 @@
boards)))) boards))))
(defn config-fill-board-defaults (defn config-fill-board-defaults
;; TODO: must have check that if board is default, it's enabled, if it's not, give some big fat warning
;; that users must always specify board, maybe change the error?
"Fills every enabled board with default config values" "Fills every enabled board with default config values"
[config] [config]
(let [defaults (:boards-defaults config)] (let [defaults (:boards-defaults config)]

View file

@ -72,6 +72,7 @@
(let [config (conf/get-some-config (:config options))] (let [config (conf/get-some-config (:config options))]
;; TODO: probably refactor to use separate config.clj file when validation will be added ;; TODO: probably refactor to use separate config.clj file when validation will be added
;; Init the few globals we have ;; Init the few globals we have
;; TODO: this all needs to go in separate function so it doesnt have to duplicated in repl-main
(reset! conf/GLOBAL-CONFIG config) (reset! conf/GLOBAL-CONFIG config)
(reset! feed/boards-enabled-cache (set (keys (get config :boards-enabled)))) (reset! feed/boards-enabled-cache (set (keys (get config :boards-enabled))))
(reset! watcher/chod-threads-cache (watcher/generate-chod-cache-structure config)) (reset! watcher/chod-threads-cache (watcher/generate-chod-cache-structure config))
@ -83,6 +84,12 @@
(defn repl-main (defn repl-main
"Development entry point" "Development entry point"
[] []
(let [config (conf/get-some-config nil)]
;; TODO: probably refactor to use separate config.clj file when validation will be added
;; Init the few globals we have
(reset! conf/GLOBAL-CONFIG config)
(reset! feed/boards-enabled-cache (set (keys (get config :boards-enabled))))
(reset! watcher/chod-threads-cache (watcher/generate-chod-cache-structure config)))
(jetty/run-jetty (rp/wrap-params #'feed/http-handler) (jetty/run-jetty (rp/wrap-params #'feed/http-handler)
{:port (:port conf/CONFIG-DEFAULT) {:port (:port conf/CONFIG-DEFAULT)
;; Dont block REPL thread ;; Dont block REPL thread

View file

@ -64,14 +64,13 @@
(let [filterable (select-keys query-string (let [filterable (select-keys query-string
(keys known-filter-map))] (keys known-filter-map))]
(ut/fkmap (fn [k v] (ut/fkmap (fn [k v]
{(get known-filter-map k) v}) {(get known-filter-map k) (ut/vectorize v)})
filterable))) filterable)))
(defn filter-chod-posts (defn filter-chod-posts
"Return list of all threads with equal or higher ChoD than requested "Return list of all threads with equal or higher ChoD than requested
;;resume
READS FROM GLOBALS: watcher.time-of-cache" READS FROM GLOBALS: watcher.time-of-cache"
[query-vec chod-treshold repeat? board-cache] [filters chod-treshold repeat? board-cache]
(let [{time-of-generation :time (let [{time-of-generation :time
cache :data} board-cache cache :data} board-cache
guid-fn (case repeat? guid-fn (case repeat?
@ -82,25 +81,21 @@
cache)) cache))
;; So we don't have to search thru everything we have cached ;; So we don't have to search thru everything we have cached
needed-cache-part (subvec cache cache-start-index) needed-cache-part (subvec cache cache-start-index)
;; Here we gonna run fmap but not really actuall-matches (keep (fn [thread]
actuall-matches (keep (fn [t] (some
(let [title (:title t)] (fn [fun]
;; Todo: Man, wouldn't it be cool to know which querry matched the thread? (when (fun thread (get filters fun))
;; Would be so much easier for user to figure out why is it showing thread))
;; and it would solve the problem of super long titles (or OPs instead of titles) (keys filters)))
(when (some (fn [querry]
(s/includes? (s/lower-case title) (s/lower-case querry)))
query-vec)
t)))
(reverse needed-cache-part))] (reverse needed-cache-part))]
;; Finally generate and append GUIDs ;; Finally generate and append GUIDs
(map guid-fn actuall-matches))) (map guid-fn actuall-matches)))
(defn thread-to-rss-item (defn thread-to-rss-item
"Converts cached thread item to feed item which can be serialized into RSS" "Converts cached thread item to feed item which can be serialized into RSS"
[t host board] [t host]
(let [link-url (s/replace host "{threadnum}" (str (:no t)))] ;Hardcode emergency bugfix (let [link-url (s/replace host "{threadnum}" (str (:no t)))]
{:title (format "%.2f%% - %s" (:chod t) (:title t)) ;TODO: Generate link from the target somehow, or just include it from API response {:title (format "%.2f%% - %s" (:chod t) (:title t))
;; :url link-url <- this is supposed to be for images according to: https://cyber.harvard.edu/rss/rss.html ;; :url link-url <- this is supposed to be for images according to: https://cyber.harvard.edu/rss/rss.html
:description (format "The thread: '%s' has %.2f%% chance of dying" (:title t) (:chod t)) :description (format "The thread: '%s' has %.2f%% chance of dying" (:title t) (:chod t))
:link link-url :link link-url
@ -108,9 +103,8 @@
(defn generate-feed (defn generate-feed
"Generates feed from matching items" "Generates feed from matching items"
[query-vec chod-treshold repeat? cache board-config self-link] [filters chod-treshold repeat? cache board-config self-link]
(let [items (filter-chod-posts query-vec chod-treshold repeat? cache) (let [items (filter-chod-posts filters chod-treshold repeat? cache)
served-filename (get @conf/GLOBAL-CONFIG :served-filename)
head {:title (str "RSS Thread watcher v" conf/VERSION) head {:title (str "RSS Thread watcher v" conf/VERSION)
;; :link is the homepage of the channel ;; :link is the homepage of the channel
:link (get @conf/GLOBAL-CONFIG :homepage) :link (get @conf/GLOBAL-CONFIG :homepage)
@ -141,22 +135,17 @@
query :query-string query :query-string
scheme :scheme scheme :scheme
server-name :server-name} rqst server-name :server-name} rqst
filters (make-filters prms) filters (make-filters prms f/known-filters)
;; qrs (prms "q") ;; BUG if local fileserver not running -> FileNotFound exception is thrown and it fucks up the feed generation
self-uri (str (s/replace-first scheme ":" "") ;; Should be handled because wrong config and thus url generation could do the same
self-uri (str (s/replace-first scheme ":" "") ;
"://" server-name uri "?" query) "://" server-name uri "?" query)
;; queries (if (vector? qrs) qrs [qrs]) ; to always return vector
real-chod (if-let [ch (or (and (vector? chod)
(first chod))
chod)]
(try ;If we can't parse number from chod, use default 94
(if (or (vector? chod)
;; TODO: Do we seriously parse this twice?
(<= (Integer/parseInt chod) 60)) ; Never accept chod lower than 60 TODO: don't hardcode this
60 (Integer/parseInt chod))
(catch Exception e
94)))
board-config (get-in @conf/GLOBAL-CONFIG [:boards-enabled board]) board-config (get-in @conf/GLOBAL-CONFIG [:boards-enabled board])
real-chod (try (max (Integer/parseInt (or (and (vector? chod)
(first chod))
chod)) 60) ;HARDCODED CHoD
(catch Exception _
(get board-config :default-chod)))
cache @watcher/chod-threads-cache] cache @watcher/chod-threads-cache]
(println "\n\nRCVD: " rqst) (println "\n\nRCVD: " rqst)
;; (println rqst) ;; (println rqst)
@ -193,7 +182,7 @@
;; There shouldn't be any problems with this mime type but if there are ;; There shouldn't be any problems with this mime type but if there are
;; replace with "text/xml", or even better, get RSS reader that is not utter shit ;; replace with "text/xml", or even better, get RSS reader that is not utter shit
:header {"Content-Type" "application/rss+xml"} :header {"Content-Type" "application/rss+xml"}
:body (generate-feed queries real-chod repeat? (watcher/get-thread-data board @conf/GLOBAL-CONFIG) board-config self-uri)}) :body (generate-feed filters real-chod repeat? (watcher/get-thread-data board @conf/GLOBAL-CONFIG) board-config self-uri)})
(catch Exception e (catch Exception e
;; Ex-info has been crafted to match HTTP response body so we can send it ;; Ex-info has been crafted to match HTTP response body so we can send it
(if-let [caught (ex-data e)] (if-let [caught (ex-data e)]

View file

@ -20,15 +20,15 @@
(defn case-sensitive-filter (defn case-sensitive-filter
"Returns true if string [s] is matched by any query. It's case insensitive" "Returns true if string [s] is matched by any query. It's case insensitive"
[s queries] [{:keys [title]} queries]
(some (fn [querry] (some (fn [querry]
(cs/includes? s querry)) (cs/includes? title querry))
queries)) queries))
(defn case-insensitive-filter (defn case-insensitive-filter
"Returns true if string [s] is case-matched by query" "Returns true if string [s] is case-matched by query"
[s queries] [{:keys [title]} queries]
(case-sensitive-filter (cs/lower-case s) (map cs/lower-case queries))) (case-sensitive-filter {:title (cs/lower-case title)} (map cs/lower-case queries)))
(def known-filters (def known-filters
{"Q" case-sensitive-filter {"Q" case-sensitive-filter

View file

@ -74,6 +74,7 @@
{k (map-apply-defaults conf-val default-val)} {k (map-apply-defaults conf-val default-val)}
{k (nil?-else conf-val default-val)}))))) {k (nil?-else conf-val default-val)})))))
;; This is a shitty version of reduce-kv
(defn fmap (defn fmap
"Applies function [f] to every key and value in map [m] "Applies function [f] to every key and value in map [m]
Function signature should be (f [key value]). Function signature should be (f [key value]).