Added: Set URL base from command line, set some of the parameters using environment variables. See #710.
Fixed: Results from nzb.su were reported as passworded (not my fault)
Fixed: Prevent hit limit reset times from being greater than 23. See #690.
Added: Option to define expiry of remember-me cookie. See #676.
Fixed: Log update would be executed even after leaving log page. See #677.
Changed: Properly report when API hit limit reached at omgwtfnzbs.org.
Fixed: Dereferer not used for NZB details links. See #653.
Fixed: Sabnzbd default category priority not applied for uploads. See #666
Changed: Renamed "umask" to "chmod". See #659
Fixed: Query generation for manual TV searches in combination with raw search engines would fail. See #651
Added: Option to set umask of log files. See #579
Fixed: Some indexers don't like colons (":") in their queries so I remove them. There shouldn't be any queries that get less specific by that. See #645
Fixed: NZBIndex queries were configured to hide crossposts which doesn't make sense. Thanks to doofy666 to making me aware.
Fixed: At least one problem introduced with 0.2.221/0.0.222: Downloading NZBs failed. See #639. Please note that it's possible that downloads for results that were found between updating to 0.2.221/0.2.222 and updating to this version might not work.
Fixed: Results with special characters in titles would cause error (mostly torrents). See #634.
Changed: Reverted the database code yet again. I can't get it to work for everybody. I give up. I hope to release a new major version of Hydra in fall that won't have this problem
Changed: Use an "empty" login page that does not expose any data.
Fixed: Some mobile layout issues. Thanks to nemchik. See #629.
Fixed: In some rare cases where indexers were selected for a search but could not execute it the search would crash. See #633.
Fixed: Jackett/Cardigann results with special characters in the title would break the search. For now they're skipped, working on a better solution. See #634.
Fixed: nzb.su has a limiting mechanism which stops the caps check from working. Added a pause for this indexer to keep it from blocking Hydra. Note: If you do many concurrent results it will probably still result in blocks but there's not much I can do about that. See #635.
Changed: I updated the database library and changed the handling again. This will hopefully solve some issues that some users still had but might impact performance. I haven't found a satisfactory solution yet. Please let me know if you experience any problems.
Added: Config option to define trailing words which will be removed from titles. Similar to "Obfuscated" and for languages, now it's just generic. Thanks to judhat2 for the list. See #604.
Added: SSL certificates will be verified by default. See #624
Added: Option to provide an SSL CA file. See #623.
Fixed: Don't convert required/forbidden regex to lowercase. See #616.
Fixed: API key generation from GUI contained many zeroes when called from Chrome. See #619.
Fixed: Use two concurrent connections when checking indexer capabilities. See #606.
Fixed: Allow nonnummeric episodes for internal TV search.. See #607.
Fixed: Previous versions were not compatible with Windows versions previous to 10... So much stuff to be careful about. See #602.
Fixed: Unable to manually shutdown or restart on windows version in some cases. See #597.
Fixed: Indexer download shares was calculated wrong. See #588.
Fixed: Windows release 0.2.214/215 did not save settings. Sorry about that -.-
Fixed: Online help didn't work.
Fixed: Connection check with SABnzbd 2.x would still fail...
Fixed: Windows service would not start with Windows 10 Creator's Update. If you're affected you'll need to uninstall the service and install it again. See #530.
Fixed: Connection check with SABnzbd 2.0RC3 would fail.
Changed: Log error when SSL could not be imported instead of exception. See #576.
Fixed: Search history empty. See #574.
Fixed: "loadLimitOnRandom" not set for new installations. See #571.
Added: Poor man's load balancing. For each indexer you can define a number x and for every API call that indexer will be picked with a chance of 1/x. For example with a value of 3 it will on average be picked on every third API call (if it meets all other restrictions). This should allow to distribute access to indexers with a low API hit limit over a day. It's up to you to find sensible values.
Changed: Rewrote the caps check error handling some more because this causes problems and questions very often. Added a message saying that you can set the IDs manually and which links to the Wiki.
Fixed: Database error when calling --help. See #563.
Fixed: Validation of URL base setting allowed a trailing slash. See #569.
Fixed: Adding new users would in rare cases cause an incomplete configuration.
Fixed: Hydra would crash trying to log the used port in some cases. See #559.
Added: The fallback mechanism added in 0.2.202 would only activate when zero results were found. You can now configure to search indexers without results even if some results were found (by other indexers). Thanks to alamei for providing the code. See #540
Added: Option to hide details, comments & NZB links and indexer selection box for certain users. This affects the GUI only; users can theoretically still "hack" them but I consider that an acceptable risk...
Added: Download a bunch of NZBs as ZIP. See #550.
Fixed: Moving backup related files would fail if moved between filesystems. Thanks to RomRider for the code.
Fixed: Prevent exception when logging failed git output. See #555.
Added: Properly log or open browser to ipv6 address. See #554.
Changed: Added option to keep "obfuscated" in NzbGeek titles. See #553.
Changed: Respect min/max age/size when provided in API calls. This is not specified in the newznab API standards but might be helpful for some. See #551.
Added: Option to fallback to a query based search if no results were found for an ID based search. Note that this will increase API hits, potentially by a lot, because it will repeat the search e.g. for every search for a movie that's not yet released. Hydra will also only search the indexers that were searched before, so if the indexer didn't support the ID in the first place and query generation is disabled it won't be search. See #540
Changed: Reformat pubdates provided by indexers. Hopefully fixes #489.
Changed: Cleaned up logging of rejection reasons.
Fixed: Repeating a movie search from the search history would use the wrong search mode.
Fixed: Use randomly generated API key instead of predefined. Thought that was already in there, oops...
Fixed: Predefined indexers were missing config variable on new installs. See #548.
Fixed: Hopefully fix database error for good.
Fixed: Forgot to migrate database, resulting in an error upon NZB download. See #543.
Fixed: NFO display was broken with last version. Also adapted NFO and log colors to grey theme.
Added: Download limit for indexers. When the download limit is reached for an indexer it will not be picked for searching. Please note that Hydra will never prevent downloads from happening even when the download limit is reached. This is to make sure that external tools do not disable Hydra when a requested download fails. See #247.
Fixed: Some hit limit related bugs.
Fixed: Multiple UI issues, e.g. tooltip placement, header "active section" display, etc.
Fixed: Search history next to search box didn't work with authorization. See #539.
Fixed: When running as a windows service updating would fail. For this to work you need to update manually (shutdown Hydra and overwrite all files from the latest ZIP) and change the user with which the service is started. Open the service administration, select the NZBHydra service, open its properties and in the second tab enter your windows account username and password. Restart the service. The next update should work. See #536.
Changed: Don't search when no query was entered. See #533.
Changed: Drunken Slug API endpoint is now https://api.drunkenslug.com. See #526.
Fixed: Binsearch NZB links could not be handled by NZBGet. See #527.
Changed: Removed data grid for history. Wrote a custom implementation for search and download history that allows sorting and filtering. At a later point I'll add that for search result filtering. See #479.
Changed: Improved error message while checking capabilities a tiny bit.
Fixed: Form bases login when using reverse proxies didn't work. See #523.
Fixed: If configured log IP for failed form logins. See #448.
Changed: Checking capabilities of indexers was supposed to find out which categories are supported and only enabled these. That never really worked well so for now I disabled that feature. Now by default all categories are enabled for new indexers.
Fixed: Some kinks with auth handling. If you have problems please post a detailed description of the scenario including debugging infos.
Changed: Improved auth handling a bit
Fixed: If a user is logged in only show his searches in the search history on the search page. See #519.
Fixed: Show distinct queries in search history on search page. See #515.
Added: You can access and repeat the latest 20 manual searches from a button on the search screen. See #515.
Changed: Changed start view of "History & Stats" to "Search history" because stats took a long time to be loaded before you could switch to one of the other subareas
Fixed: Restored button to repeat searches and display of metaquery for searches without query. See #515.
Fixed: Various other small bugs related to the search history and its presentation.
Fixed: Adding an indexer without using a preset was impossible. See #514.
Fixed: Update library for handling dates and extended newznab time patterns to hopefully properly parse usenet dates from indexers that change their date format for no fucking reason. See #512.
Fixed: Sometimes indexers were not picked if they didn't support an ID but it would've been possible to convert the ID to a supported one (e.g. TMDB requested and IMDB supported).
Added: Show an error if an indexer was ignored during when searching (every enabled indexer should either be picked for searching or it should be mentioned why it wasn't). See #511.
Added: Show tooltips on search result icons and give the icons a bit more space. See #506.
Added: Button to scroll to bottom of log file and option to update log file automatically. See #497 and #498.
Added: Options to control when a new log file is started (depending on size or time), how many to keep and if a new file should be started on startup. See #496.
Added: Data grid for search history which allows filtering and sorting. Not fully implemented yet but will be used for search and download history, with proper category / date filtering. See #479.
Changed: Removed womble :-(
Fixed: Show better error message when connection test to indexer failed. See #500.
Fixed: NZBGeek and forbidden words don't play well together. See #493.
Fixed: Error in ignoring posters introduced in 0.2.180.
Added: Ignore results by certain posters and/or from certain groups. See #478.
Added: Set user agent per indexer. Few will ever need this. See #482.
Added: Allow deleting of preconfigured indexers. Of course they can be added later again. See #480.
Fixed: Manually checking indexer capabilities didn't work. See #477.
Added: Scripts for windows release to easily create a windows service.
Fixed: Indexer username / password were not used for connection and capabilities check. See #465.
Misc: I've only had 70 survey responses so far compared to nearly 300 on the first survey. Please take part, if possible.
Fixed: Duplicate detection age and size were ignored. See #463.
Fixed: When opening the browser on startup is enabled the configured host IP is used instead of 127.0.0.1. Apparently there are cases where that doesn't work for some reason.
Misc: I made a new survey with Google Forms (see below). If you chose to ignore the last one you will not be asked again. Otherwise a new popup will ask you take part in the new survey. Either way you can visit it here: Google Forms survey. If you're privacy minded you can open that link in a private browser window. I also made sure that new users will not immediately be asked to take part in the survey. Instead the popup will appear after three days.
Changed: Age of search results is shown more precisely (minutes, hours, days). See #460.
Added: Send indexer name, host and score with API search results. Useful for tools and scripts handling them. See #455.
Added: More presets, provided by judhat2. See #461
Fixed: Required and forbidden words were not honored when using "Load more". See #459.
Fixed: "tuple index out of range" error in stat generation. See #456.
Changed: Removed log window from windows tray tool because it caused some trouble and I didn't want to debug it... See #453.
Misc: Thank you everybody for answering the survey, I really appreciate it. Unfortunately I'm an idiot and SurveyMonkey only shows me the first 100 results for the free plan and I'm not paying 40 euros to upgrade. So... if you were one of the first 100 your voice will be heard, otherwise not - so sorry about that. Anyway, I'll collect what I got and especially check your comments. I'll make a collection issue on GitHub to answer all questions and requests that I can see. I might make a new survey when I found a provider that doesn't want heaps of money...
Added: The search page will show a popup asking you to answer a few questions in a survey. The survey is completely anonymous (I disabled IP address logging) and only about Hydra. It would help me a lot if you could participate, it will take only a few minutes. If you'd rather participate from another computer or want the URL for some reason without having to visit the page: https://www.surveymonkey.com/r/HWXLCHM Thank you.
Added: When "Shutdown" to restart is enabled Hydra will send exit code 6 when shutting down. Together with the correct configuration this allows linux service managers to recognize that the instance did not crash and should not be restarted. Thank you, MarMed. #446
Fixed: Don't display indexers as disabled if they're not. Also log a bit more infos when and why a disabler gets disabled and when it will be reenabled. See #447.
Fixed: What was supposed to fix duplicate indexers in the index statuses made it worse. Oops. See #445.
Fixed: Downloading NZBs via API failed with 0.2.171. See #443.
Fixed: Make sure indexer statuses are not duplicated. Also fixed some kinks in the handling of indexer failures and the reenabling of indexers via GUI. See #366.
Added: Option to shutdown Hydra when it's supposed to restart (either after an update or as requested by the user). This could help users who run Hydra with a service manager on Linux which restarts the instance when it closes and thus produces a second instance. See #442.
Changed: Updated BeautifulSoup library which will hopefully fix conflicts with newer linux installations.
Changed: Search result IDs are now derived from their actual indexer search results. That means that when a search result is found, deleted from the database and found again it will have the same ID both times. This should ensure that CP doesn't find the same result after 7+ days and thinks it's a new one. All existing search result IDs are invalid. The updating process might take a bit longer than usual.
Fixed: Downloader icons wouldn't update according to status.
Added: Button to force an update from GitHub.
Fixed: Unique results share buggy from 0.2.169.
Added: Allow providing username and password for socks proxy. Be aware that only direct calls from Hydra will go through proxies. Calls to git for example will not. See #430.
Added: Allow not defining a category when sending NZBs to a downloader. In that case the downloader will decide which category to use. See #434.
Added: Show stats for a certain time span. See #431.
Added: Option to truncate database and log file. Requested by a user running the synology package where database and log files are not available easily.
Added: New "dark" theme which is basically just a darker grey theme. Supplied by /u/SabreWolF9. Grey theme is now default.
Changed: API keys are not mandatory for indexers. Apparently there are som which don't require one or even throw an error when one is provided. Whatever. See #427
Changed: Some indexers do support more search types or IDs than they say in their caps. So search capabilities of an indexer (searching by IMDB ID, TVDB ID, etc.) are now always determined by "brute force" instead of relying on the data provided in the caps. That means that the caps check will take a lot more time (up to 6 queries) but be more precise. See #433.
Fixed: Calls to /details/ would not use dereferer. See #434.
Fixed: Database was not properly shut down.
Changed: I lost access to my old bitcoin wallet. I created a new one and updated the readme. As I don't have any means of "converting" money to bitcoin and have some services that I pay with Bitcoin I changed my policy and am now open to receive donations ;-)
Fixed: Details link was extracted wrong for some newznab indexers. Will only affect new results not already in the database.
Fixed: Was possible to select the search type (Internal vs API) for Jackett / Cardigann indexers. Torrents are only supported for internal searches.
Fixed: Debug code cached API searches resulting in no new searches being executed.
Fixed: Search would work when an indexer was not picked. See #418 (again).
Fixed: Search would work when Binsearch or NZBIndex returned no results. See #418.
Added: Display in search results and log how many search results were rejected for which reason. See #417.
Changed: Remove old omgwtfnzbs implementation now that they support a proper newznab compatible API. Shoutout to their team for improving the indexer and helping me support it. See #416.
Changed: Increased default max size for TV results to 4500 to include MULTI 1080p BluRay rips which tend to be decadently large.
Changed: Rename omgwtfnzbs.org to omgwtfnzbs and change references in database to keep stats. See #414.
Fixed: Max age was not respected in internal movie and TV searches and never used for newznab indexer requests. See #412.
Fixed: Error while searching with windows binaries. See #413
Fixed: Windows binaries wouldn't start up because of library inconsistencies. See #409.
Changed: Properly use newznab indexer data to check if an NFO is available. See #329.
Fixed: Results from indexers reporting no comments would show comment icon. See #411.
Fixed: "Include duplicates" state would not be remembered properly and switching it was wonky.
Fixed: Expanding duplicates fucked up the layout (since 0.2.154).
Fixed: Sorting by grabs works with indexers which don't report the number of grabs. Results from those will be considered as having 0 grabs. See #407.
Fixed: Size unavailable for Cardigann results, errors where multiple results with the same GUID were returned by Jackett/Cardigann. See #410.
Added: This changelog will now contain links to the associated issues on GitHub. If you have questions or feedback you're welcome to comment there.
Added: Hydra will try to determine what backend a newznab compatible indexer is running (newznab, nntmux or nzedb) and depending on that use the correct values for NFO retrieval and excluding words. Because a search needs to be executed for this analysis it will not be done automatically, so please open the settings for each newznab compatible indexer and click "Check capabilities". Sorry about that. See #376 and #329.
Added: Results in a category for which an indexer is not enabled will be rejected (when searching without specified categories). See #353.
Changed: New omgwtfnzbs host. Existing settings will be migrated. See #409.
Fixed: Sorting by grabs. See #407.
Fixed: Stat times weren't localized. See #378.
Added: Torrent search compatible with Cardigann.
Changed: (Hopefully) better algorithm to get details links for NZBs.
Changed: Redesigned results details (again). Removed number of files and comments, replaced text with icons, show comment icon if comments are available. Known issue: No tooltip for fuzzy NFO links.
Nothing important.
Fixed: Getting debug infos might fail with reverse proxy.
Added: Store Ebook search data in the database. Redesigned search history.
Added: Support for dereferer. Enabled by default.
Added: Support for HTTP(S) proxy.
Changed: Removed some 6box indexers from the presets.
Fixed: Restart after update wouldn't work.
Fixed: Database migration of 0.2.149 would not store new database version meaning the migration might've been done on every startup.
Added: Show number of files, grabs and comments where indexers provide this information.
Fixed: Config for removing duplicates for API results was left over.
Fixed: When you enable or disable an indexer in the config, save and go to the search page the indexer's new state is proper represented (i.e. it's not hidden or shown without you having to reload the page).
Added: Restore settings from backup via GUI. Please note that this only supports backup ZIPs created with this version or later.
Added: Display number of rejected results in search results view and new button to load all available results.
Added: Enhanced stat calculation: * Show the average daily number of API accesses for each indexers * Show the average number of unique results each indexer contributed in a search (will only work for searches made from this point on). Also added some text explaining that ;-) * Show average number of downloads and searches per day of week and hour of day * Added some fancy charts
Changed: Removed option that decided if duplicates should be removed for API calls, this is now always enabled.
Changed: Changed and rewrote database handling. Should (mostly) prevent the notorious "database is locked" errors.
Fixed: Downloads would always be displayed as "API" in the history.
Fixed: NZBIndex result titles would be stripped of all spaces.
Fixed: Indexer statuses box wouldn't remember state.
Fixed: Enabling and disabling indexers would not be properly broadcasted to all areas of the code.
Fixed: Some color inconsistencies.
Fixed: Better handle NZBClub returning an empty page.
Fixed: Better handle parsing errors in general.
Changed: Rewrote search result database handling in the hope of fixing this schlamassel.
Fixed: Database indexes were never created. This is the reason why saving search results got so slow after some time and, after 0.2.140, search results would be duplicated. The migration code needs to delete search results from the database. If you get errors in Sonarr or CP that downloads failed this is the reason. Sorry about that.
Fixed: Login would fail in some cases.
Fixed: After last version queries would be cut if pressing return too fast.
Added: Filter results directly on the search results page. Just edit the query or change/add min/max age/size and the already loaded results are filtered on the fly.
Fixed: Remove debug code.
Changed: Rewrote some of the search result presentation logic which improves rendering time a bit (still slow, though).
Added/Fixed: Button to invert selection is back and should work as expected.
Added: Shift-click support for the result checkboxes. Click one result, shift click another and all shown displayed checkboses between them will be set to the new value of the first clicked checkbox.
Changed: Replaced collapse icons with + and - which I find easier do understand / recognize.
Changed: Hopefully improved the performance of writing results to the database. Let me know what happens.
Added: Log how long a search took.
Added: Some performance related debug loggings.
Added: Button to invert indexer selection on earch page.
Fixed: Indexer statuses box would not store state properly.
Fixed: Womble would not search at all.
Added: Log a warning if no indexers were picked and provide the reasons.
Fixed: Properly follow configuration if required words / regexes should be used for internal or external searches. Added such a switch to the searching config.
Added: Required and forbidden regex support. Separate values for now, so you can either just use comma separated values and / or complex regexes.
Added: Option to remove trailing language from newznab results. Some indexers add the language to the result title, preventing proper duplicate detection. Enabled by default.
Changed: "Average results shares" statistic now only considers specific searches, so only those where you actually searched for somethic specific, using a query or an ID. This should results in more helpful values. Note that they're still not precise. When you do a search with only one indexer enable its share will be 100%. If you search with two indexers and the second is offline, then the first one's share is 100% (and the second one's not counted).
Fixed: Sonarr search was broken after last update. Sorry...
Added: Support for daily shows.
Fixed: Migration from config version 22 didn't work so all users with that version would get errors. Sorry about that.
Changed: Added a small section in the online help to explain indexer priorities.
Fixed: Better error handling with testing indexer caps.
Fixed: Fallback if indexer was not properly added to the database.
Fixed: Fallback if indexer capabilities settings are not properly set for some reason.
Fixed: Allow empty username and password for sabnzbd.
Changed: Remove "-Obfuscated" from NZBGeek results for better title pairing.
Changed: Better error message if an indexer API access could not be saved
Changed: Drastically improved processing time. Duplicate detection is about 60% faster. Searches with a lot of new results (not already in the database) are up to 7 times faster (which only shows how bad it was before).
Added: Set "no referrer" tag.
Added: API hit limit rolling reset. The indexer doesn't reset the hits at a certain point but you're not allowed to make more than x limits in a 24 hour window. This is used by default.
Changed: Hit limit reset time is defined as hour of day now instead of time of day.
Fixed: Categories not set for new indexers not added by selecting a preset.
Fixed: When searching TV shows or movies from the GUI using autocomplete don't use the title as query.
Added: Include newznab categories in API results where possible.
Fixed: Use proper age threshold for duplicate detection when poster and group are unknown.
Changed: Tweaked (and hopefully improved) duplicate detection.
Added: Include search query in title of search results page.
Fixed: Duplicates were not displayed properly.
Fixed: Problem with categories (I hope).
Added: Newznab indexers will now be checked for their supported categories and by default only be enabled for those. For those categories that are not standardized by newznab (e.g. Anime, Ebooks, etc.) Hydra will try to determine the used category identifier for each indexer. Please note that unfortunately there is no automatic process to do this for indexers that were already added. Please click the "Check types" button for every indexer once and then save. Sorry about that.
Added: Anime category and anizb indexer.
Added: When selecting a category on the search page only those indexers that are enabled for that category will be displayed.
Fixed: It looks like I found a solution to the long-standing "database is locked" issue. If true, the fix was embarrassingly, hilariously simple and obvious.
Fixed: Entries in stats and indexer statuses are sorted (either better to worse or by name).
Fixed: API key wasn't displayed in omgwtf config.
Fixed: Exception when API call with categories and offset > 0 was executed.
Added: Basic torrent support. You need to run and configure Jackett. Add trackers using the "Jackett" preset in the configuration. They will be only used for internal searches. Torrent clients are not supported and will never be. You can search manually and download the torrent files. That should be enough for most.
Added: Limit indexers to certain categories. Makes most sense when you have added specialized torrent trackers and don't want to use them for categories in which they don't have content anyway.
Changed: Required and forbidden words are now handled a bit differently. If a word contains a dot or a dash (".", "-") then the word will be searched for anywhere in a result title. If it does not contain a dot or a dash then all the words in the title will be compared with the word. Example: Forbidding "abc" will dismiss the result "abc.def" or "abc def" but allow "abcdef" or "ab-cdef". Forbidding "ab-c" will dismis "ab-cdef". This way you'll be able, for example, to require release groups like "EA" without allowing a result like "Peachy-BOSS" and on the other hand forbidding "WEB-DL" without letting through "Spider.web-Dl.ist". I think this is the approach that will handle the most common cases.
Changed: Required words are now searched on a word basis instead of full-text, meaning that at least one of the required words needs to be a word in a result's title, not just be present anywhere in the title.
Fixed: "All" category wasn't available in selection box after having selected another category.
Added: SOCKS proxy support by sanderjo
Fixed: Return correct NZBGeek details links.
Fixed: Lxml wouldn't load even if installed properly.
Fixed: Download of debug infos wouldn't work in firefox.
Fixed: Make sure iconCssClass is set.
Fixed: Category mapping wouldn't work with omgwtf.
Fixed: URL base was not included in NZB links when external URL was not set. Now the local IP address, configured port and scheme are used. Tools on your computer (and inside your network) should be able to use the generated links. If you need to send links to tools outside your network you have to set the external URL.
Fixed: Excluding words with "--" in the search field didn't work.
Fixed: Total number of results was not put into API search result if offset was 0.
Changed: Rewrote auth handling. Unfortunately form based auth only works when calls from the GUI are done, as soon as you call any function from outside (e.g. CP) the token header is missing and you will be asked for credentials using basic auth. Now, when you enter them, hydra will accept them even with form based auth enabled. I might change that behavior in the future but for now it stands.
Fixed: Repeating search from history wouldn't work (in multiple ways).
Fixed: ID based search from the GUI didn't work.
Fixed: Forbidden and ignored words not displayed in config.
Fixed: Download of bugreport would not work with auth type "Form".
Fixed: Error with API search and categories due to stupid mistake but caused by fucking Python 2.7.
Fixed: Set appropriate categories when API searches are done ("Movies" for movie search, etc.)
Fixed: Stats and config invisible with no auth enabled.
Added: Several changes in the handling and configuration of categories.
-
New category section in the config. Define required and forbidden words for every category and if they should be applied for external and/or internal searches. This allows you to finetune results for CP or Sonarr, for example.
-
Define newznab categories for every Hydra category. These categories will be used when you search with that category from the GUI. That way you can control if foreign movies should be included in movie searches, for example.
-
Decide for every category if you want to ignore results from it for external and/or internal searches.
Added: Online help in the configuration.
Added: Improved test coverage and started to write more integration tests. Should somewhat decrease chance of breaking bugs, at least for API. I still don't have frontend tests.
Changed: Login and logout functionality. The navigation only shows links to sections the current user is allowed to visit. Click the login button in the upper right to login. Click it again to logout. If you use basic auth and logged out make sure to close the browser so that the auth header is removed.
Changed: Don't require auth for NZB details links when form based auth is enabled. Should prevent troubles with viewing details from external tools like CP and the security risks are negligible.
Fixed: Don't display an NZB download as failed if it was only redirected (then we don't know if it was successful).
Fixed: Error where indexers tab in config could not be opened.
Fixed: When migration of old config fails keep the config instead of overwriting it. Provide feedback and stop program.
Fixed: Support for custom downloader icons was broken.
Changed: Using ! to exclude words when querying newznab indexers. This should be compatible with more indexers and provide the wanted results. Thanks to judhat2 for the research and tests.
Fixed: Fallback to pubdate when usenet date cannot be parsed.
Added: Return category in API search results.
Fixed: If log file name is provided in command line show that in GUI instead of the configured one.
Added: Show log in scrollable area.
Added: Comic category. Searches in newznab comic category if available, otherwise in general ebook category or just for the query in case of raw search engines.
Added: Link to indexer details page for entries in download history.
Fixed: Selecting search results from duplicates / same name rows wouldn't work.
Fixed: Search would ask for auth even if access wasn't restricted.
Fixed: Downloader default category wouldn't be used.
Changed: Provide a bit better info when auth fails.
Changed: Provide better feedback when the connection test to an indexer fails.
Fixed: Form based auth wouldn't work with subdirectories on reverse proxies.
Fixed: Indexer status box state wasn't remembered.
Fixed: Changelog would either show too much or not enough...
Fixed: Search box would disappear when showing search results.
Fixed: Sensitive downloader data would be sent to non-admins. Sorry.
Fixed: Changelog would be shown empty.
Added: Decide if you want to authenticate using HTTP basic auth or a login form. If you have users configured please make sure that everything is in order.
Added: SimplyNZBs added to the presets.
Fixed: Displayed times in stats were all sorts of wrong.
Probably not fixed but less likely: Database is locked.
Fixed: Details link would point to "http://api.indexer.com/details/..." and not work.
Fixed: When adding multiple indexers one after the other the data from the old indexer would be still there.
Added: Option if IP addresses should be logged. Failed logins will still be logged with the used IP address.
Fixed: /details links wouldn't work.
Fixed: Bug report info generating should work again.
Fixed even more: Allow special characters in NZBGet username and password.
Changed: Jumped to version 0.2.91 because I'm an idiot and the version check did a string comparison, so 0.2.91 was the next minor version that would be found as an update. Also fixed some bugs in the changelog retrieval.
Added: Make sure that indexers' and downloaders' names are unique.
Added: Remember state of indexer status box on search results page and remember sorting.
Added: Popover name of downloader and allow to select an icon from http://fontawesome.io/icons/ instead of the default icon.
Fixed: NZB link should work with URL base.
Fixed: Details link for nZEDb based indexers should link to details pages instead of info pages.
Fixed: Stats and history should display times in proper timezone.
Yeah, well, 0.2.7 changed a lot and broke a lot. Sorry for the problems, I'm trying to get most of them fixed in the coming days.
Fixed: omgwtfnzbs.org results are parsed correctly and will not be skipped.
Fixed: If results are incomplete and not added to the database they're not included in the results either.
Fixed: Config did not like downloader passwords with certain special characters.
Fixed: Clicking on downloader button would fail.
Changed: Try to prevent "database is locked error" or at least get better logs.
Added: Support multiple downloaders.
Added: When the connection test for a downloader or indexer fails you can decide if you want to add it anyway.
Changed: Don't wipe the search field when changing the category. If you already entered something and you select a movie or TV category the autocomplete function is triggered automatically.
Added: Hide disabled indexers in statistics. This only affects the visibility, they're still calculated. If an indexer was never enabled this might skew the statistics, but for now I can't do better.
Fixed: Statistics for downloads and results share were not generated.
People are getting more database locked errors. If you have found a realiable way of reproducing it please let me know. As a test I configured a database timeout. Send me a message if it causes troubles.
Fixed: Binsearch downloads would fail after binsearch changed their API.
Fixed: Downloading and NFO and details retrieval via API didn't work since 0.2.0.
Changed: Made sure that the GUID of the returned results is unique ("nzbhydrasearchresult" + internal ID). This may cause some issues with tools that already retrieved results and stored them, but nothing too serious.
Fixed: API search results would contain no GUID so sonarr thought they were all the same. This is a quick fix to get things back to running, will improve this later.
Bump version to make sure update is executed. Jump to 0.2.x seems to have caused some troubles.
Fixed: Constraint errors handling search results.
Fixed: Small problem with last update and empty details links.
Changed: Moved minor version to 0.2.0. I would like to say that's because I feel Hydra is more mature now but it's actually because my versioning algorithm thought that 0.0.1a103 < 0.0.1a99...
Changed: From now on hydra will save all found search results to the database for a certain amount of time. This allows easier communication between server and client and some features that I'm planning. This also means that for the first time I need to do a bit of data migration. The download history should remain intact, but there's the chance that in some circumstances the migration fails. Sorry...!
Fixed: Indexers couldn't be deleted in config.
Fixed: NZBGeek queries contained a superfluous "--".
Fixed: Bugreport wouldn't work.
Added: When leaving the config page with unsaved changes ask if the user wants to save, discard or stay on the config page.
Fixed: Some problems with resetting the indexer config.
Fixed: Layout issues in indexer config. Still not perfect but better.
Fixed: Show API reset time for indexers in UTC and make sure UTC is used.
Added: Option to always show duplicates in the search results (see Config -> Searching -> Result processing - > Always show duplicates).
Added: Show indexer priority (renamed from "Score") in main indexer view. Sort indexers by priority and allow to change it right there.
Fixed: Newznab indexers were shown twice on search page since last update.
Fixed: ipinfo lib not found in binary release
Changed: I redesigned the indexer configuration to look like the one in sonarr. It is a lot less cluttered now. Also connection tests are done automatically and for new indexers the capabilities test is executed automatically. Let me know what you think.
Fixed: Properly recognize empty result pages from binsearch and don't show stack trace if the HTML could not be parsed.
Added: Support for details and getnfo via API.
Added: Log IP of NZB downloader. Thanks to sanderjo.
Added: Make errors in config dialog better visible
Fixed: Bug in max age limit for results.
Fixed: Color in select list using dark theme.
Added: Ubuntu upstart scripts and better daemon control/logging. Thanks to nikcub.
Added: Option to set git executable used for updates.
Fixed: Don't try to exclude words with dots, dashes or spaces in queries if the indexer doesn't support it. Indexers that don't follow newznab standards (-- prefix for exclusions) will still not work properly, for now.
Changed: Minor changes in dark theme.
Added: Dark theme. Feedback appreciated.
Added: Show NZB downloads per indexer in stats.
Added: HTTP auth for indexers (rarely needed).
Fixed: Re-enable indexer on status page.
Added: Feature to limit the number of maximum API hits for an indexer in 24 hours.
Added: Show proper page titles.
Added: Show titles for ID based searches in history (if already known).
Fixed: TypeError when using OMGWTF.
Fixed: Unable to load details page.
Changed: Try to solve database locked error.
Added: Add NZB Finder to presets.
Added: Support JSON output for API search results.
Added: Globally define words of which at least one needs to be contained in displayed results.
Added: Link to TVDB pages from search history for TVRage ID based searches.
Fixed: Error when searching movies with titles containing special characters using the frontend.
Fixed: Error in search because I wrote code while too tired and checked in without testing. Shame on me.
Fixed: Binsearch results where age could not be parsed caused problems, will be ignored.
Fixed: Leave settings.cfg untouched if an error occurrs instead of writing an empty file.
Fixed: Don't crash whole app if exception in search thread is thrown.
Fixed: Make sure that searches are not executed with empty converted search IDs. This would sometimes cause false positives being returned.
Added: Support for book searches via API. Because a lot of newznab indexers don't support book queries I decided to use query generation instead, meaning I do a raw query with the supplied author and/or title in the ebook categories. This might return some/a lot of false positives, but it's better than no results at all.
Changed: All raw indexers (Binsearch, NZBClub and NZBIndex) are not only enabled for internal searches only. They return too much crap that tools using the API will not be able to handle properly as they expect correctly indexed results.
Added: Prefix terms in query with "--" to exclude them. Works in addition to global ignored words.
Fixed: Skip indexers if query generation is enabled, they don't support an ID and the retrieval of the title for the query generation failed.
Fixed: Properly cache retrieved titles for TV or movie IDs.
Fixed: Unable to add user without admin rights.
Fixed: Exclude results violating age or size filters directly in indexer queries where possible and if not then during result processing, not in the GUI.
Fixed: Problems with validation and general usage of authorization config.
Fixed: Don't show update footer for users without admin rights.
Added: New "Bugreport" tab in the "System" section which gives some advice and provides functions to download anonymized versions of the settings and log which you can post.
Added: Automatically create backup of database and settings before updating. New "Backup" tab in "System" section to create a backup and download existing backup files.
Added: Global setting for maximum age of results.
Added: Errors in the web client will be logged by the server.
Fixed: Properly recognize if wrong schema was supplied for downloader host.
Fixed: If the "Ignore words" settings ended with a comma all search results would be ignored.
Fixed: Use relative base href instead of absolute. Should be compatible with IIS reverse proxy now. No need to preserve the host anymore.
Fixed: Loading of more results would fail.
Fixed: Binsearch results would sometimes have the wrong age.
Changed: Config help texts are now left aligned.
Fixed: Binsearch would sometimes return duplicate results.
Changed: Indexer statuses on search results page is minimizable and minimized by default.
Changed: Checkbox to invert selection temporarily disabled because of a bug. Will reenable it when fixed.
Fixed: Searching TV shows by season/episode via GUI didn't work.
Fixed: Testing connection of downloader in config didn't work.
Fixed: Don't provide cert. You need to use your own or better, use a reverse proxy.
Changed: Threaded server is now activated by default. Improves page loading times and allows parallel searches (yay).
Fixed: Startup would fail without existing settings.cfg.
Fixed: SSL verification failures on Qnap.
Changed: Reworked caching of assets. Simplifies development process and makes updates a lot smaller.
Fixed bug where buttons for newznab indexer tests wouldn't work
Fixed bug where newznab indexers' settings would be incomplete and cause an error when searching.
Fixed valiidation in config so that the indexer timeout is optional.
Fixed bug where sending NZBs to downloader wouldn't work. Removed docker from readme. Will be moved to wiki. Open NZB details in new tab/window.
Fixed bug where NFO retrieval didn't work because of a JS error. My bad.
Fixed bug where searching didn't work because of a JS error. Whoops.
Rewrote and simplified code for settings which finally allows using an unlimited amount of newznab indexers, along with better GUI handling of those. This affected basically every feature in the program, so from experience I'd say I fucked up something which I didn't find during testing, so please let me know ;-)
Replaced simple users with multi-user system. Add as many users as you want and control if they're allowed to use basic features, see the stats and/or have admin rights. All future searches and downloads will be logged with these users.
Removed caching of search results because it didn't really work and nobody uses it anyway. Therefore removed all cache related settings.
Added validation for most settings in config. It's still possible to make mistakes but... just don't be stupid ;-)
Fixed bug where searches with empty query parameters would be sent to indexers.
Improved handling of failed logins.
Increase timeout for sabNZBd and add logging.
Fixed bug with newznab search type detection where only a couple of results would be shown in some cases.
Fixed bug where sending links to downloaders would fail with enabled auth.
Completely rewrote duplicate detection. Fixes an ugly bug, should take 2/3 of the time and easier to fix or expand in the future.
Added argument switches for PID file and log file location.
When an indexer wasn't searched (e.g. because it doesn't support any of the search types) a message will be shown and the search is not considered unsuccessful.
Use proper caching so that the assets should only be reloaded when they've actually changed (and then actually reload). Should make page loading faster on slow upstream servers and solve problems with outdated assets.
Moved about, updates, log and control sections to their own "System" tab (like sonar ;-)).
Added version history to updates tab.
Still getting used to writing the change log so I might often forget it for a while.
Fixed a bug where duplicate detection would ironically cause duplicates which caused some weird bugs in the system. Was a pain in the ass to debug and fix.
Added an option to look at this (the changelog) before updating.
Removed "direct" NZB access type. Programs will always need to contact NZB Hydra to get their NZBs.
First version with changelog
Split settings for base URL and external URL in two. Added option to use local URL for search results.
Show notification when update is available.
Prepared for windows release. Expect it in the next week or so.
Spotweb results should now be parsed properly.
Last version without changelog