* Use the request context
The code uses context.Background() in a flow where there is a
http.Request. Use the requests context instead.
* Use a true context in the plugin example
Let AddTag/RemoveTag take a context and use that context throughout
the example.
* Avoid the use of context.Background
Prefer context.TODO over context.Background deep in the call chain.
This marks the site as something which we need to context-handle
later, and also makes it clear to the reader that the context is
sort-of temporary in the code base.
While here, be consistent in handling the `act` variable in each
branch of the if .. { .. } .. check.
* Prefer context.TODO over context.Background
For the different scraping operations here, there is a context
higher up the call chain, which we ought to use. Mark the call-sites
as TODO for now, so we can come back later on a sweep of which parts
can be context-lifted.
* Thread context upwards
Initialization requires context for transactions. Thread the context
upward the call chain.
At the intialization call, add a context.TODO since we can't break this
yet. The singleton assumption prevents us from pulling it up into main for
now.
* make tasks context-aware
Change the task interface to understand contexts.
Pass the context down in some of the branches where it is needed.
* Make QueryStashBoxScene context-aware
This call naturally sits inside the request-context. Use it.
* Introduce a context in the JS plugin code
This allows us to use a context for HTTP calls inside the system.
Mark the context with a TODO at top level for now.
* Nitpick error formatting
Use %v rather than %s for error interfaces.
Do not begin an error strong with a capital letter.
* Avoid the use of http.Get in FFMPEG download chain
Since http.Get has no context, it isn't possible to break out or have
policy induced. The call will block until the GET completes. Rewrite
to use a http Request and provide a context.
Thread the context through the call chain for now. provide
context.TODO() at the top level of the initialization chain.
* Make getRemoteCDPWSAddress aware of contexts
Eliminate a call to http.Get and replace it with a context-aware
variant.
Push the context upwards in the call chain, but plug it before the
scraper interface so we don't have to rewrite said interface yet.
Plugged with context.TODO()
* Scraper: make the getImage function context-aware
Use a context, and pass it upwards. Plug it with context.TODO()
up the chain before the rewrite gets too much out of hand for now.
Minor tweaks along the way, remove a call to context.Background()
deep in the call chain.
* Make NOTIFY request context-aware
The call sits inside a Request-handler. So it's natural to use the
requests context as the context for the outgoing HTTP request.
* Use a context in the url scraper code
We are sitting in code which has a context, so utilize it for the
request as well.
* Use a context when checking versions
When we check the version of stash on Github, use a context. Thread
the context up to the initialization routine of the HTTP/GraphQL
server and plug it with a context.TODO() for now.
This paves the way for providing a context to the HTTP server code in a
future patch.
* Make utils func ReadImage context-aware
In almost all of the cases, there is a context in the call chain which
is a natural use. This is true for all the GraphQL mutations.
The exception is in task_stash_box_tag, so plug that task with
context.TODO() for now.
* Make stash-box get context-aware
Thread a context through the call chain until we hit the Client API.
Plug it with context.TODO() there for now.
* Enable the noctx linter
The code is now free of any uncontexted HTTP request. This means we
pass the noctx linter, and we can enable it in the code base.
* Add collation to directory listings. Closes#1806
Introduce a new `locale` arg to the `Query.directory` field. Set "en"
as the default for the field for backward compatibility. Use the given
locale, sending it through a language matcher, and use `x/text` as the
collation engine for the matched language.
Augment the file `ListDirs` call to optionally take a Collator. If the
Collator is given, sort file listings according to the collators rules.
While here, document the GraphQL schema a bit more.
Add matchers by looking at the current front-end locales, and make sure
each of these occur in the matcher list.
* Language matcher touchups
* Avoid having `en-US` twice.
* Introduce `en-AU`.
* Pass IgnoreCase and Numeric collation
Allow the collator to be configured with options. Pass the options
IgnoreCase and Numeric to the collator.
* Replace error assertions with Go 1.13 style
Use `errors.As(..)` over type assertions. This enables better use of
wrapped errors in the future, and lets us pass some errorlint checks
in the process.
The rewrite is entirely mechanical, and uses a standard idiom for
doing so.
* Use Go 1.13's errors.Is(..)
Rather than directly checking for error equality, use errors.Is(..).
This protects against error wrapping issues in the future.
Even though something like sql.ErrNoRows doesn't need the wrapping, do
so anyway, for the sake of consistency throughout the code base.
The change almost lets us pass the `errorlint` Go checker except for
a missing case in `js.go` which is to be handled separately; it isn't
mechanical, like these changes are.
* Remove goconst
goconst isn't a useful linter in many cases, because it's false positive
rate is high. It's 100% for the current code base.
* Avoid direct comparison of errors in recover()
Assert that we are catching an error from recover(). If we are,
check that the error caught matches errStop.
* Enable the "errorlint" checker
Configure the checker to avoid checking for errorf wraps. These are
often false positives since the suggestion is to blanket wrap errors
with %w, and that exposes the underlying API which you might not want
to do.
The other warnings are good however, and with the current patch stack,
the code base passes all these checks as well.
* Configure rowserrcheck
The project uses sqlx. Configure rowserrcheck to include said package.
* Mechanically rewrite a large set of errors
Mechanically search for errors that look like
fmt.Errorf("...%s", err.Error())
and rewrite those into
fmt.Errorf("...%v", err)
The `fmt` package is error-aware and knows how to call err.Error()
itself.
The rationale is that this is more idiomatic Go; it paves the
way for using error wrapping later with %w in some sites.
This patch only addresses the entirely mechanical rewriting caught by
a project-side search/replace. There are more individual sites not
addressed by this patch.
Reduce allocations. Don't create intermediary arrays which we then
consume right after. Manually fuse the arrays and decode straight into
the sum instead.
Furthermore, don't invoke a Reader, but carve out the locations via a
loop, directly.
These two changes taken together speeds up oshash computations by a
factor of 10 according to the benchmark tests. The main reason for
this change is a much lowered memory allocation rate which in turn
improves GC pressure.
While here, add a benchmark for oshash computations and use it for
testing the performance.
* Refactor scraper structures
* Move matching code into new package
* Add autotag scraper
* Always check first letter of auto-tag names
* Account for nulls
Co-authored-by: Kermie <kermie@isinthe.house>
* Add security against publicly exposed services
* Add trusted proxies setting, validate proxy chain against internet access
* Validate chain on local proxies too
* Move authentication handler to separate file
* Add startup check and log if tripwire is active
Co-authored-by: WithoutPants <53250216+WithoutPants@users.noreply.github.com>
* Enable safe linters
Enable the linters dogsled, rowserrcheck, and sqlclosecheck.
These report no errors currently in the code base.
Enable misspell.
Misspell finds two spelling mistakes in comments, which are fixed by the
patch as well.
Add and sort linters which are relatively
safe to add over time. Comment them out for now.
* Close the response body
If we can get a HTTP response, it has a body which ought to be closed.
By doing so, we avoid potentially leaking connections.
* Enable the exportloopref linter
There are two places in the code with these warnings. Fix them while
enabling the linter.
* Remove redundant types in tests
If a slice already determines the type, the inner type declaration is
redundant. Remove the inner declarations.
* Mark autotag test cases as parallel
Autotag test cases is by far the outlier when it comes to test time.
While go test runs test cases in parallel,
it doesn't do so inside a given package, unless one marks the test cases
as parallel.
This change provides a significant speedup on a 8-core machine for test
runs.
* Add API support for filtering tags by parent / children
* Add parent & child tags filters for tags to UI
* Add API support for filtering tags by parent / child count
* Add parent & child count filters for tags to UI
* Update db generator
* Add missing build tag
* Add unit tests
Co-authored-by: WithoutPants <53250216+WithoutPants@users.noreply.github.com>
* Fix all revive warnings in the code base
All of these are of the form
```
var Identifier Type = Expr
```
where the `Type` is known from the output of `Expr` and can be omitted
as a result.
* Handle unchecked errors
* Remove new-from-rev
Since the project passes all linter checks now, including older
revisions, we can remove new-from-rev. While here, reorder the linter
config file, and move the enabled linters up and settings down.
* Fix failing test cases
Studio & Performer export tests use local time rather than UTC. This
fixes the test cases so integration test
passes.
* Add movies and movie_count properties to Performer type
Extend the GraphQL API to allow getting the movies and movie count by
performer.
* Add movies count to performer card
* Add movies and movie_count properties to Studio type
Extend the GraphQL API to allow getting the movies and movie count by
studio.
* Add movies count to studio card
The io/ioutil package has been deprecated as of Go 1.16, see
https://golang.org/doc/go1.16#ioutil. This commit replaces the existing
io/ioutil functions with their new definitions in io and os packages.
Signed-off-by: Eng Zer Jun <engzerjun@gmail.com>
* Support getting scenes on movies in the API
* Make movie card visually consistent with scene etc
* Add date
* Add synopsis
* Show scene count with hover listing the scenes
* Move scene index to button
* Move scene number to own section
Co-authored-by: WithoutPants <53250216+WithoutPants@users.noreply.github.com>
* Log 3 unchecked errors
Rather than ignore errors, log them at
the WARNING log level.
The server has been functioning without these, so assume they are not at
the ERROR level.
* Log errors in concurrency test
If we can't initialize the configuration, treat the test as a failure.
* Undo the errcheck on configurations for now.
* Handle unchecked errors in pkg/manager
* Resolve unchecked errors
* Handle DLNA/DMS unchecked errors
* Handle error checking in concurrency test
Generalize config initialization, so we can initialize a configuration
without writing it to disk.
Use this in the test case, since otherwise the test fails to write.
* Handle the remaining unchecked errors
* Heed gosimple in update test
* Use one-line if-initializer statements
While here, fix a wrong variable capture error.
* testing.T doesn't support %w
use %v instead which is supported.
* Remove unused query builder functions
The Int/String criterion handler functions are now generalized.
Thus, there's no need to keep these functions around anymore.
* Mark filterBuilder.addRecursiveWith nolint
The function is useful in the future and no other refactors are looking
nice.
Keep the function around, but tell the linter to ignore it.
* Remove utils.Btoi
There are no users of this utility function
* Return error on scan failure
If we fail to scan the row when looking for the
unique checksum index, then report the error upwards.
* Fix comments on exported functions
* Fix typos
* Fix startup error
* Improve image scanning performance and thumbnail generation
* Add vips-tools to build image
* Add option to write generated thumbnails to disk
* Fallback to image if thumbnail generation fails
Co-authored-by: WithoutPants <53250216+WithoutPants@users.noreply.github.com>
Rather than passing a pointer to a waitgroup into task.Start(..)
functions, handle the waitgroup.Done() at the callsite.
This makes waitgroup handling local to its definition rather than it
being spread out over multiple files. Tasks now simply execute, and
the policy of waiting on them is handled by the caller.
* Support subpaths when serving stash through a reverse proxy
* Add README documentation
Co-authored-by: WithoutPants <53250216+WithoutPants@users.noreply.github.com>
* Add indexes for path and checksum to images
The scenes table has unique indexes/constraints on path and checksum
colums. The images table doesn't, which doesn't really make sense, as
scanning uses these colums extensively which warrents an index, and both
should be unique as well.
Adding these indexes thus heavily improves the scanning tasks
performance. On a database containing 4700 images a (re)scan of those
4700 files, which thus shouldn't do anything, took 1.2 seconds, with the
indexes added this only takes 0.4 seconds. Taking the same test on a
generated database containing 4M images + the actual 4700 images took 26
minutes for a rescan, and with the index existing also only takes 0.4
seconds.
* Add images.checksum unique constraint in code with fallback
Work around the issue where in some cases duplicate images (/checksums
on images) might exist. This as discussed in #1740 by creating the index
on startup and in case of an error logging the duplicates. This so the
users where this scenario exists can correct the database (by searching
on the logged checksum(s) and removing the duplicates) and after a
restart the unique index / constraint will still be created. In case
when creating the unique index fails a "normal" / non-unique index is
created as surrogate so the user will still get the performance benefit
(for example during scanning) without being forced to remove the
duplicates and restart beforehand. This surrogate is also automatically
cleaned up after the unique index is succesfully created.
NORMAL sync is safe when using WAL journaliing.
It cuts the amount of sync calls to the disk, resulting in faster write
operation. On power loss, the database will perhaps lose some ongoing
commits, but that is all.
The expectation is that if the database lives on the same disk as the
stash, this could help performance quite a bit under heavier operation.
* Avoid redundant logging in migrations
Return the error and let the caller handle the logging of the error if
needed.
While here, defer m.Close() to the function boundary.
* Treat errors as values
Use %v rather than %s and pass the errors directly.
* Generate a wrapped error on stat-failure
* Log 3 unchecked errors
Rather than ignore errors, log them at
the WARNING log level.
The server has been functioning without these, so assume they are not at
the ERROR level.
* Propagate errors upward
Failure in path generation was ignored. Propagate the errors upward the
call stack, so it can be handled at the level of orchestration.
* Warn on errors
Log errors rather than quenching them.
Errors are logged at the Warn-level for now.
* Check error when creating test databases
Use the builtin log package and stop the program fatally on error.
* Add warnings to uncheck task errors
Focus on the task system in a single commit, logging unchecked
errors as warnings.
* Warn-on-error in API routes
Look through the API routes, and make sure errors are being logged if
they occur. Prefer the Warn-log-level because none of these has proven
to be fatal in the system up until now.
* Propagate error when adding Util API
* Propagate error on adding util API
* Return unhandled error
* JS log API: propagate and log errors
* JS Plugins: log GQL addition failures.
* Warn on failure to write to stdin
* Warn on failure to stop task
* Wrap viper.BindEnv
The current viper code only errors if no name is provided, so it should
never fail. Rewrite the code flow to factor through a panic-function.
This removes error warnings from this part of the code.
* Log errors in concurrency test
If we can't initialize the configuration, treat the test as a failure.
* Warn on errors in configuration code
* Plug an unchecked error in gallery zip walking
* Warn on screenshot serving failure
* Warn on encoder screenshot failure
* Warn on errors in path-handling code
* Undo the errcheck on configurations for now.
* Use one-line initializers where applicable
rather than using
err := f()
if err!= nil { ..
prefer the shorter
if err := f(); err != nil { ..
If f() isn't too long of a name, or wraps a function with a body.
* Execute Gallery.Create.Post plugin hook during scan
Fix issue where Gallery.Create.Post hook is not executed when a new
gallery is created during scan, when the gallery is created based on the
folder.
* Fix Gallery.Create.Post hook being invoked in transaction
Invoke the Gallery.Create.Post hook during zip scan after the
transaction has been committed. This is necessary to allow the plugin to
access the gallery (using GraphQL API). Otherwise the API obviously uses
a different database transaction which can't find the gallery as it
isn't committed yet.
* Fix logs from scraper and plugins not being shown in UI
Using `logger.` in the logger package to write logs is "incorrect". This
as the package contains a variable named `logger` which contains the
logrus instance. So instead of the log line being handled by the custom
log implementation / wrapper which makes sure the lines are shown in the
UI as well, it's written to logrus directly meaning the wrapper is
skipped.
This "issue" is obviously triggered because in any other place
`logger.X` can be used and it will used the custom logger package /
wrapper which works fine.
* Add plugin / scraper name to logging output
Indicate which plugin / scraper wrote a log message by including its
name to the `[Scrape]` prefix.
* Add missing addLogItem call
* Generate screenshot images for markers
In some scenarios it might not be possible to use the preview video or
image of markers, i.e. when only static images are supported like in
Kodi. So generate a static screenshot as well.
* Make generating animated and static image optional for markers
* Use screenshot for markers when preview type is set to static image
* Add migration to create studio aliases table
* Refactor studioQueryBuilder.Query to use filterBuilder
* Expand GraphQL API with aliases support for studio
* Add aliases support for studios to the UI
* List aliases in details panel
* Allow editing aliases in edit panel
* Add 'aliases' filter when searching
* Find studios by alias in filter / select
* Add auto-tagging based on studio aliases
* Support studio aliases for filename parsing
* Support importing and exporting of studio aliases
* Search for studio alias as well during scraping
* Add migration script for tag relations table
* Expand hierarchical filter features
Expand the features of the hierarchical multi input filter with support
for using a relations table, which only has parent_id and child_id
columns, and support adding an additional intermediate table to join on,
for example for scenes and tags which are linked by the scenes_tags
table as well.
* Add hierarchical filtering for tags
* Add hierarchical tags support to scene markers
Refactor filtering of scene markers to filterBuilder and in the process
add support for hierarchical tags as well.
* List parent and child tags on tag details page
* Support setting parent and child tags
Add support for setting parent and child tags during tag creation and
tag updates.
* Validate no loops are created in tags hierarchy
* Update tag merging to support tag hierarcy
* Add unit tests for tags.EnsureUniqueHierarchy
* Fix applying recursive to with clause
The SQL `RECURSIVE` of a `WITH` clause only needs to be applied once,
imediately after the `WITH`. So this fixes the query building to do just
that, automatically applying the `RECURSIVE` keyword when any added with
clause is added as recursive.
* Rename hierarchical root id column
* Rewrite hierarchical filtering for performance
Completely rewrite the hierarchical filtering to optimize for
performance. Doing the recursive query in combination with a complex
query seems to break SQLite optimizing some things which means that the
recursive part might be 2,5 second slower than adding a static
`VALUES()` list. This is mostly noticable in case of the tag hierarchy
where setting an exclusion with any depth (or depth: all) being applied
has this performance impact of 2,5 second. "Include" also suffered this
issue, but some rewritten query by joining in the *_tags table in one
pass and applying a `WHERE x IS NOT NULL` filter did seem to optimize
that case. But that optimization isn't applied to the `IS NULL` filter
of "exclude". Running a simple query beforehand to get all (recursive)
items and then applying them to the query doesn't have this performance
penalty.
* Remove UI references to child studios and tags
* Add parents to tag export
* Support importing of parent relationship for tags
* Assign stable ids to parent / child badges
* Silence Apollo warning on parents/children fields on tags
Silence warning triggered by Apollo GraphQL by explicitly instructing it
to use the incoming parents/children values. By default it already does
this, but it triggers a warning as it might be unintended that it uses
the incoming values (instead of for example merging both arrays).
Setting merge to false still applies the same behaviour (use only
incoming values) but silences the warning as it's explicitly configured
to work like this.
* Rework detecting unique tag hierarchy
Completely rework the unique tag hierarchy to detect invalid hierarchies
for which a tag is "added in the middle". So when there are tags A <- B
and A <- C, you could previously edit tag B and add tag C as a sub tag
without it being noticed as parent A being applied twice (to tag C).
While afterwards saving tag C would fail as tag A was applied as parent
twice. The updated code correctly detects this scenario as well.
Furthermore the error messaging has been reworked a bit and the message
now mentions both the direct parent / sub tag as well as the tag which
would results in the error. So in aboves example it would now show the
message that tag C can't be applied because tag A already is a parent.
* Update relations on cached tags when needed
Update the relations on cached tags when a tag is created / updated /
deleted so these always reflect the correct state. Otherwise (re)opening
a tag might still show the old relations untill the page is fully
reloaded or the list is navigated. But this obviously is strange when
you for example have tag A, create or update tag B to have a relation to
tag A, and from tags B page click through to tag A and it doesn't show
that it is linked to tag B.
