git-annex/Command/MetaData.hs

186 lines
5.9 KiB
Haskell
Raw Normal View History

{- git-annex command
-
- Copyright 2014-2016 Joey Hess <id@joeyh.name>
-
- Licensed under the GNU AGPL version 3 or higher.
-}
module Command.MetaData where
import Command
import Annex.MetaData
import Annex.VectorClock
import Logs.MetaData
import Annex.WorkTree
import Messages.JSON (JSONActionItem(..), AddJSONActionItemFields(..))
import Types.Messages
import Utility.Aeson
import Limit
import qualified Data.Set as S
import qualified Data.Map as M
import qualified Data.Text as T
import qualified Data.ByteString.Char8 as B8
import qualified Data.ByteString.Lazy.UTF8 as BU
import Control.Concurrent
cmd :: Command
cmd = withGlobalOptions [jsonOptions, annexedMatchingOptions] $
command "metadata" SectionMetaData
"sets or gets metadata of a file"
paramPaths (seek <$$> optParser)
data MetaDataOptions = MetaDataOptions
{ forFiles :: CmdParams
, getSet :: GetSet
, keyOptions :: Maybe KeyOptions
, batchOption :: BatchMode
}
2014-03-17 19:26:18 +00:00
data GetSet = Get MetaField | GetAll | Set [ModMeta]
optParser :: CmdParamsDesc -> Parser MetaDataOptions
optParser desc = MetaDataOptions
<$> cmdParams desc
<*> ((Get <$> getopt) <|> (Set <$> some modopts) <|> pure GetAll)
<*> optional parseKeyOptions
<*> parseBatchOption False
where
getopt = option (eitherReader (mkMetaField . T.pack))
( long "get" <> short 'g' <> metavar paramField
<> help "get single metadata field"
)
modopts = option (eitherReader parseModMeta)
( long "set" <> short 's' <> metavar "FIELD[+-]=VALUE"
<> help "set or unset metadata value"
)
<|> (AddMeta tagMetaField . toMetaValue . encodeBS <$> strOption
( long "tag" <> short 't' <> metavar "TAG"
<> help "set a tag"
))
<|> (DelMeta tagMetaField . Just . toMetaValue . encodeBS <$> strOption
( long "untag" <> short 'u' <> metavar "TAG"
<> help "remove a tag"
))
<|> option (eitherReader (\f -> DelMeta <$> mkMetaField (T.pack f) <*> pure Nothing))
( long "remove" <> short 'r' <> metavar "FIELD"
<> help "remove all values of a field"
)
<|> flag' DelAllMeta
( long "remove-all"
<> help "remove all metadata"
)
seek :: MetaDataOptions -> CommandSeek
seek o = case batchOption o of
NoBatch -> do
c <- currentVectorClock
let ww = WarnUnmatchLsFiles
let seeker = AnnexedFileSeeker
{ startAction = start c o
, checkContentPresent = Nothing
, usesLocationLog = False
}
let seekaction = case getSet o of
Get _ -> withFilesInGitAnnex ww
GetAll -> withFilesInGitAnnex ww
Set _ -> withFilesInGitAnnexNonRecursive ww
"Not recursively setting metadata. Use --force to do that."
withKeyOptions (keyOptions o) False seeker
(commandAction . startKeys c o)
(seekaction seeker)
=<< workTreeItems ww (forFiles o)
Batch fmt -> withMessageState $ \s -> case outputType s of
JSONOutput _ -> ifM limited
( giveup "combining --batch with file matching options is not currently supported"
, batchOnly (keyOptions o) (forFiles o) $
batchInput fmt parseJSONInput
(commandAction . batchCommandStart . startBatch)
)
_ -> giveup "--batch is currently only supported in --json mode"
deal better with clock skew situations, using vector clocks * Deal with clock skew, both forwards and backwards, when logging information to the git-annex branch. * GIT_ANNEX_VECTOR_CLOCK can now be set to a fixed value (eg 1) rather than needing to be advanced each time a new change is made. * Misuse of GIT_ANNEX_VECTOR_CLOCK will no longer confuse git-annex. When changing a file in the git-annex branch, the vector clock to use is now determined by first looking at the current time (or GIT_ANNEX_VECTOR_CLOCK when set), and comparing it to the newest vector clock already in use in that file. If a newer time stamp was already in use, advance it forward by a second instead. When the clock is set to a time in the past, this avoids logging with an old timestamp, which would risk that log line later being ignored in favor of "newer" line that is really not newer. When a log entry has been made with a clock that was set far ahead in the future, this avoids newer information being logged with an older timestamp and so being ignored in favor of that future-timestamped information. Once all clocks get fixed, this will result in the vector clocks being incremented, until finally enough time has passed that time gets back ahead of the vector clock value, and then it will return to usual operation. (This latter situation is not ideal, but it seems the best that can be done. The issue with it is, since all writers will be incrementing the last vector clock they saw, there's no way to tell when one writer made a write significantly later in time than another, so the earlier write might arbitrarily be picked when merging. This problem is why git-annex uses timestamps in the first place, rather than pure vector clocks.) Advancing forward by 1 second is somewhat arbitrary. setDead advances a timestamp by just 1 picosecond, and the vector clock could too. But then it would interfere with setDead, which wants to be overrulled by any change. So it could use 2 picoseconds or something, but that seems weird. It could just as well advance it forward by a minute or whatever, but then it would be harder for real time to catch up with the vector clock when forward clock slew had happened. A complication is that many log files contain several different peices of information, and it may be best to only use vector clocks for the same peice of information. For example, a key's location log file contains InfoPresent/InfoMissing for each UUID, and it only looks at the vector clocks for the UUID that is being changed, and not other UUIDs. Although exactly where the dividing line is can be hard to determine. Consider metadata logs, where a field "tag" can have multiple values set at different times. Should it advance forward past the last tag? Probably. What about when a different field is set, should it look at the clocks of other fields? Perhaps not, but currently it does, and this does not seems like it will cause any problems. Another one I'm not entirely sure about is the export log, which is keyed by (fromuuid, touuid). So if multiple repos are exporting to the same remote, different vector clocks can be used for that remote. It looks like that's probably ok, because it does not try to determine what order things occurred when there was an export conflict. Sponsored-by: Jochen Bartl on Patreon
2021-08-03 20:45:20 +00:00
start :: CandidateVectorClock -> MetaDataOptions -> SeekInput -> RawFilePath -> Key -> CommandStart
start c o si file k = startKeys c o (si, k, mkActionItem (k, afile))
where
afile = AssociatedFile (Just file)
2014-03-17 19:26:18 +00:00
deal better with clock skew situations, using vector clocks * Deal with clock skew, both forwards and backwards, when logging information to the git-annex branch. * GIT_ANNEX_VECTOR_CLOCK can now be set to a fixed value (eg 1) rather than needing to be advanced each time a new change is made. * Misuse of GIT_ANNEX_VECTOR_CLOCK will no longer confuse git-annex. When changing a file in the git-annex branch, the vector clock to use is now determined by first looking at the current time (or GIT_ANNEX_VECTOR_CLOCK when set), and comparing it to the newest vector clock already in use in that file. If a newer time stamp was already in use, advance it forward by a second instead. When the clock is set to a time in the past, this avoids logging with an old timestamp, which would risk that log line later being ignored in favor of "newer" line that is really not newer. When a log entry has been made with a clock that was set far ahead in the future, this avoids newer information being logged with an older timestamp and so being ignored in favor of that future-timestamped information. Once all clocks get fixed, this will result in the vector clocks being incremented, until finally enough time has passed that time gets back ahead of the vector clock value, and then it will return to usual operation. (This latter situation is not ideal, but it seems the best that can be done. The issue with it is, since all writers will be incrementing the last vector clock they saw, there's no way to tell when one writer made a write significantly later in time than another, so the earlier write might arbitrarily be picked when merging. This problem is why git-annex uses timestamps in the first place, rather than pure vector clocks.) Advancing forward by 1 second is somewhat arbitrary. setDead advances a timestamp by just 1 picosecond, and the vector clock could too. But then it would interfere with setDead, which wants to be overrulled by any change. So it could use 2 picoseconds or something, but that seems weird. It could just as well advance it forward by a minute or whatever, but then it would be harder for real time to catch up with the vector clock when forward clock slew had happened. A complication is that many log files contain several different peices of information, and it may be best to only use vector clocks for the same peice of information. For example, a key's location log file contains InfoPresent/InfoMissing for each UUID, and it only looks at the vector clocks for the UUID that is being changed, and not other UUIDs. Although exactly where the dividing line is can be hard to determine. Consider metadata logs, where a field "tag" can have multiple values set at different times. Should it advance forward past the last tag? Probably. What about when a different field is set, should it look at the clocks of other fields? Perhaps not, but currently it does, and this does not seems like it will cause any problems. Another one I'm not entirely sure about is the export log, which is keyed by (fromuuid, touuid). So if multiple repos are exporting to the same remote, different vector clocks can be used for that remote. It looks like that's probably ok, because it does not try to determine what order things occurred when there was an export conflict. Sponsored-by: Jochen Bartl on Patreon
2021-08-03 20:45:20 +00:00
startKeys :: CandidateVectorClock -> MetaDataOptions -> (SeekInput, Key, ActionItem) -> CommandStart
startKeys c o (si, k, ai) = case getSet o of
Get f -> startingCustomOutput k $ do
l <- S.toList . currentMetaDataValues f <$> getCurrentMetaData k
liftIO $ forM_ l $
B8.putStrLn . fromMetaValue
make CommandStart return a StartMessage The goal is to be able to run CommandStart in the main thread when -J is used, rather than unncessarily passing it off to a worker thread, which incurs overhead that is signficant when the CommandStart is going to quickly decide to stop. To do that, the message it displays needs to be displayed in the worker thread, after the CommandStart has run. Also, the change will mean that CommandStart will no longer necessarily run with the same Annex state as CommandPerform. While its docs already said it should avoid modifying Annex state, I audited all the CommandStart code as part of the conversion. (Note that CommandSeek already sometimes runs with a different Annex state, and that has not been a source of any problems, so I am not too worried that this change will lead to breakage going forward.) The only modification of Annex state I found was it calling allowMessages in some Commands that default to noMessages. Dealt with that by adding a startCustomOutput and a startingUsualMessages. This lets a command start with noMessages and then select the output it wants for each CommandStart. One bit of breakage: onlyActionOn has been removed from commands that used it. The plan is that, since a StartMessage contains an ActionItem, when a Key can be extracted from that, the parallel job runner can run onlyActionOn' automatically. Then commands won't need to worry about this detail. Future work. Otherwise, this was a fairly straightforward process of making each CommandStart compile again. Hopefully other behavior changes were mostly avoided. In a few cases, a command had a CommandStart that called a CommandPerform that then called showStart multiple times. I have collapsed those down to a single start action. The main command to perhaps suffer from it is Command.Direct, which used to show a start for each file, and no longer does. Another minor behavior change is that some commands used showStart before, but had an associated file and a Key available, so were changed to ShowStart with an ActionItemAssociatedFile. That will not change the normal output or behavior, but --json output will now include the key. This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
next $ return True
_ -> starting "metadata" ai si $
make CommandStart return a StartMessage The goal is to be able to run CommandStart in the main thread when -J is used, rather than unncessarily passing it off to a worker thread, which incurs overhead that is signficant when the CommandStart is going to quickly decide to stop. To do that, the message it displays needs to be displayed in the worker thread, after the CommandStart has run. Also, the change will mean that CommandStart will no longer necessarily run with the same Annex state as CommandPerform. While its docs already said it should avoid modifying Annex state, I audited all the CommandStart code as part of the conversion. (Note that CommandSeek already sometimes runs with a different Annex state, and that has not been a source of any problems, so I am not too worried that this change will lead to breakage going forward.) The only modification of Annex state I found was it calling allowMessages in some Commands that default to noMessages. Dealt with that by adding a startCustomOutput and a startingUsualMessages. This lets a command start with noMessages and then select the output it wants for each CommandStart. One bit of breakage: onlyActionOn has been removed from commands that used it. The plan is that, since a StartMessage contains an ActionItem, when a Key can be extracted from that, the parallel job runner can run onlyActionOn' automatically. Then commands won't need to worry about this detail. Future work. Otherwise, this was a fairly straightforward process of making each CommandStart compile again. Hopefully other behavior changes were mostly avoided. In a few cases, a command had a CommandStart that called a CommandPerform that then called showStart multiple times. I have collapsed those down to a single start action. The main command to perhaps suffer from it is Command.Direct, which used to show a start for each file, and no longer does. Another minor behavior change is that some commands used showStart before, but had an associated file and a Key available, so were changed to ShowStart with an ActionItemAssociatedFile. That will not change the normal output or behavior, but --json output will now include the key. This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
perform c o k
deal better with clock skew situations, using vector clocks * Deal with clock skew, both forwards and backwards, when logging information to the git-annex branch. * GIT_ANNEX_VECTOR_CLOCK can now be set to a fixed value (eg 1) rather than needing to be advanced each time a new change is made. * Misuse of GIT_ANNEX_VECTOR_CLOCK will no longer confuse git-annex. When changing a file in the git-annex branch, the vector clock to use is now determined by first looking at the current time (or GIT_ANNEX_VECTOR_CLOCK when set), and comparing it to the newest vector clock already in use in that file. If a newer time stamp was already in use, advance it forward by a second instead. When the clock is set to a time in the past, this avoids logging with an old timestamp, which would risk that log line later being ignored in favor of "newer" line that is really not newer. When a log entry has been made with a clock that was set far ahead in the future, this avoids newer information being logged with an older timestamp and so being ignored in favor of that future-timestamped information. Once all clocks get fixed, this will result in the vector clocks being incremented, until finally enough time has passed that time gets back ahead of the vector clock value, and then it will return to usual operation. (This latter situation is not ideal, but it seems the best that can be done. The issue with it is, since all writers will be incrementing the last vector clock they saw, there's no way to tell when one writer made a write significantly later in time than another, so the earlier write might arbitrarily be picked when merging. This problem is why git-annex uses timestamps in the first place, rather than pure vector clocks.) Advancing forward by 1 second is somewhat arbitrary. setDead advances a timestamp by just 1 picosecond, and the vector clock could too. But then it would interfere with setDead, which wants to be overrulled by any change. So it could use 2 picoseconds or something, but that seems weird. It could just as well advance it forward by a minute or whatever, but then it would be harder for real time to catch up with the vector clock when forward clock slew had happened. A complication is that many log files contain several different peices of information, and it may be best to only use vector clocks for the same peice of information. For example, a key's location log file contains InfoPresent/InfoMissing for each UUID, and it only looks at the vector clocks for the UUID that is being changed, and not other UUIDs. Although exactly where the dividing line is can be hard to determine. Consider metadata logs, where a field "tag" can have multiple values set at different times. Should it advance forward past the last tag? Probably. What about when a different field is set, should it look at the clocks of other fields? Perhaps not, but currently it does, and this does not seems like it will cause any problems. Another one I'm not entirely sure about is the export log, which is keyed by (fromuuid, touuid). So if multiple repos are exporting to the same remote, different vector clocks can be used for that remote. It looks like that's probably ok, because it does not try to determine what order things occurred when there was an export conflict. Sponsored-by: Jochen Bartl on Patreon
2021-08-03 20:45:20 +00:00
perform :: CandidateVectorClock -> MetaDataOptions -> Key -> CommandPerform
perform c o k = case getSet o of
Set ms -> do
oldm <- getCurrentMetaData k
let m = combineMetaData $ map (modMeta oldm) ms
addMetaDataClocked k m c
next $ cleanup k
_ -> next $ cleanup k
cleanup :: Key -> CommandCleanup
cleanup k = do
m <- getCurrentMetaData k
case toJSON' (AddJSONActionItemFields m) of
Object o -> maybeShowJSON $ AesonObject o
_ -> noop
showLongNote $ unlines $ concatMap showmeta $
map unwrapmeta (fromMetaData m)
return True
where
2014-02-23 17:58:16 +00:00
unwrapmeta (f, v) = (fromMetaField f, map fromMetaValue (S.toList v))
showmeta (f, vs) = map ((T.unpack f ++ "=") ++) (map decodeBS vs)
parseJSONInput :: String -> Annex (Either String (Either RawFilePath Key, MetaData))
parseJSONInput i = case eitherDecode (BU.fromString i) of
Left e -> return (Left e)
Right v -> do
let m = case itemFields v of
Nothing -> emptyMetaData
Just m' -> m'
case (itemKey v, itemFile v) of
(Just k, _) -> return $
Right (Right k, m)
(Nothing, Just f) -> do
f' <- liftIO $ relPathCwdToFile (toRawFilePath f)
return $ Right (Left f', m)
(Nothing, Nothing) -> return $
Left "JSON input is missing either file or key"
startBatch :: (SeekInput, (Either RawFilePath Key, MetaData)) -> CommandStart
startBatch (si, (i, (MetaData m))) = case i of
Left f -> do
2020-07-10 18:17:35 +00:00
mk <- lookupKey f
case mk of
2019-06-06 16:53:24 +00:00
Just k -> go k (mkActionItem (k, AssociatedFile (Just f)))
metadata --batch: Avoid crashing when a non-annexed file is input Turns out that CommandStart actions do not have their exceptions caught, which is why the giveup was causing a crash. Mostly these actions do not do very much work on their own, but it does seem possible there are other commands whose CommandStart also throws an exception. So, my first attempt at a fix was to catch those exceptions. But, --json-error-messages then causes a difficulty, because in order to output a json error message, an action needs to have been started; that sets up the json object that the error message will be included in a field of. While it would be possible to output an object with just an error field, this would be json output of a format that the user has no reason to expect, that happens only in an exceptional circumstance. That is something I have always wanted to avoid with the json output; while git-annex man pages don't document what the json looks like, the output has always been made to be self-describing. Eg, it includes "error-messages":[] even when there's no errors. With that ruled out, it doesn't seem a good idea to catch CommandStart exceptions and display the error to stderr when --json-error-messages is set. And so I don't know if it makes sense to catch exceptions from that at all. Maybe I'd have a different opinion if --json-error-messages did not exist though. So instead, output a blank line like other batch commands do. This also leaves open the possibility of implementing support for matching object with metadata --json, which would also want to output a blank line when the input didn't match. Sponsored-by: Dartmouth College's DANDI project
2021-11-01 17:38:14 +00:00
Nothing -> return Nothing
Right k -> go k (mkActionItem k)
where
go k ai = starting "metadata" ai si $ do
let o = MetaDataOptions
{ forFiles = []
, getSet = if MetaData m == emptyMetaData
then GetAll
else Set $ map mkModMeta (M.toList m)
, keyOptions = Nothing
, batchOption = NoBatch
}
t <- currentVectorClock
-- It would be bad if two batch mode changes used exactly
-- the same timestamp, since the order of adds and removals
-- of the same metadata value would then be indeterminate.
-- To guarantee that never happens, delay 1 microsecond,
-- so the timestamp will always be different. This is
-- probably less expensive than cleaner methods,
-- such as taking from a list of increasing timestamps.
liftIO $ threadDelay 1
make CommandStart return a StartMessage The goal is to be able to run CommandStart in the main thread when -J is used, rather than unncessarily passing it off to a worker thread, which incurs overhead that is signficant when the CommandStart is going to quickly decide to stop. To do that, the message it displays needs to be displayed in the worker thread, after the CommandStart has run. Also, the change will mean that CommandStart will no longer necessarily run with the same Annex state as CommandPerform. While its docs already said it should avoid modifying Annex state, I audited all the CommandStart code as part of the conversion. (Note that CommandSeek already sometimes runs with a different Annex state, and that has not been a source of any problems, so I am not too worried that this change will lead to breakage going forward.) The only modification of Annex state I found was it calling allowMessages in some Commands that default to noMessages. Dealt with that by adding a startCustomOutput and a startingUsualMessages. This lets a command start with noMessages and then select the output it wants for each CommandStart. One bit of breakage: onlyActionOn has been removed from commands that used it. The plan is that, since a StartMessage contains an ActionItem, when a Key can be extracted from that, the parallel job runner can run onlyActionOn' automatically. Then commands won't need to worry about this detail. Future work. Otherwise, this was a fairly straightforward process of making each CommandStart compile again. Hopefully other behavior changes were mostly avoided. In a few cases, a command had a CommandStart that called a CommandPerform that then called showStart multiple times. I have collapsed those down to a single start action. The main command to perhaps suffer from it is Command.Direct, which used to show a start for each file, and no longer does. Another minor behavior change is that some commands used showStart before, but had an associated file and a Key available, so were changed to ShowStart with an ActionItemAssociatedFile. That will not change the normal output or behavior, but --json output will now include the key. This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
perform t o k
mkModMeta (f, s)
| S.null s = DelMeta f Nothing
| otherwise = SetMeta f s