Table of Contents generated with DocToc

• Zredis
  • Rationale
  • Deleting From Database
  • Compiling modules
  • Cache
  • News
  • Mapping Of Redis Types To Zsh Data Structures
    • Database string keys -> Zsh hash
    • Redis hash -> Zsh hash
    • Redis set -> Zsh array
    • Redis sorted set -> Zsh hash
    • Redis list -> Zsh array
    • Redis string key -> Zsh string
• Installation
  • Zplugin
  • Antigen
  • Oh-My-Zsh
  • Zgen
• Zredis Zstyles

Zsh binary module written in C interfacing with redis database via Zshell variables mapped to keys or the whole database.

# Create a hash `HASHSET' using the standard cli tool, holding 2 keys
% redis-cli -n 3 hmset HASHSET key1 value1 \
                                key2 value2

# Bind the shell variable `hset' onto the hash `HASHSET' in the database 127.0.0.1/3
% ztie -d db/redis -a "127.0.0.1/3/HASHSET" hset

% echo ${(kv)hset}    # (kv) – keys and values of Zsh hash
key1 value1 key2 value2
% echo ${(k)hset}
key1 key2
% echo $hset          # in Zsh when accesing a hash the values are output by default
value1 value2

% ztie -d db/redis -a "127.0.0.1/3/LIST" -L list lst # Lazy binding, will create list-key on write
                                                     # -L {type}, obtains Redis type name like `zset',
                                                     # `hash', `string', etc.
% echo ${(t)lst}      # (t) – display type of Zsh variable
array-special
% lst=( ${(k)hset} )  # Copying hash keys into list
% echo $lst
key1 key2
% redis-cli -n 3 lrange LIST 0 -1   # Test the value of `LIST' using the standard cli tool
1) "key1"
2) "key2"

Building commands for redis-cli quickly becomes inadequate. For example, if copying of one hash to another one is needed, what redis-cli invocations are needed? With zredis, this task is simple:

% ztie -r -d db/redis -a "127.0.0.1/3/HASHSET1" hset1 # -r - read-only
% ztie -d db/redis -a "127.0.0.1/3/HASHSET2" hset2
% echo ${(kv)hset2}
other data
% echo ${(kv)hset1}
key1 value1 key2 value2
% hset2=( "${(kv)hset1[@]}" )
% echo ${(kv)hset2}
key1 value1 key2 value2

The "${(kv)hset1[@]}" construct guarantees that empty elements (keys or values) will be preserved, thanks to quoting and @ operator. (kv) means keys and values, alternating.

Or, for example, if one needs a large sorted set (zset), how to accomplish this with redis-cli? With zredis, one can do:

% redis-cli -n 3 zadd NEWZSET 1.0 a
% ztie -d db/redis -a "127.0.0.1/3/NEWZSET" zset
% echo ${(kv)zset}
a 1
% count=0
% for i in {a..z} {A..Z}; do (( count ++ )); zset[$i]=$count; done
% echo ${(kv)zset}
a 1 b 2 c 3 d 4 e 5 f 6 g 7 h 8 i 9 j 10 k 11 l 12 m 13 n 14 o 15 p 16 q 17 r 18 s 19 t 20 u 21 v 22 w 23 x 24 y 25 z 26 A 27 B 28 C 29 D 30 E 31 F 32 G 33 H 34 I 35 J 36 K 37 L 38 M 39 N 40 O 41 P 42 Q 43 R 44 S 45 T 46 U 47 V 48 W 49 X 50 Y 51 Z 52
% zrzset -h
Usage: zrzset {tied-param-name}
Output: $reply array, to hold elements of the sorted set
% zrzset zset
% echo $reply
a b c d e f g h i j k l m n o p q r s t u v w x y z A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Unsetting the first type of mapped variable (Zsh hash -> whole database) doesn't cause a deletion from database. If option -D is given to ztie when binding to concrete key in database, then unsets, also caused by automatic Zsh scoping actions, cause the corresponding key to be deleted. zuntie never deletes from database.

More: in Redis, removing all elements from a set, list, etc. means the same as deletion. So you can delete all datatypes except string, by doing variable=(). For string you can unset key in whole-database mapped hash: unset 'wholedb[key]'.

The Zsh modules provided by the plugin will build automatically (hiredis library is needed). You can start more than 1 shell, only the first one will be compiling. If a developer commits a new timestamp to module/RECOMPILE_REQUEST, the module will recompile (don't worry, at startup, mtime is checked first, so check for recompilation is fast). I do this when I add tested features or fixes. You can recompile the modules yourself by invoking Zsh function zredis_compile.

By default, reads are cached. If a tied variable is read for the first time, then database is accessed. For the second read there's no database access. Writes aren't cached in any way.

To clear the cache, invoke:

ztclear my_string_var       # Also for types: list, set
ztclear my_hashset_var key  # Also for types: whole-db mapping, zset

To disable the cache, pass -z ("zero-cache") option to ztie.

• 2018-12-19

  • The builtin zrpush can have the param-name argument skipped – if it's called for the second time, meaning that a new special (but writeable) parameter has been set – $zredis_last. It holds the param-name used in the 1st call and will be used in place of the {pm-name} argument. The short call is then to look like the following: zrpush {l|r} [ {val1} {val2} ... ].

• 2018-12-18

  • New builtin zrpush {l|r} {pm-name} [ {val1} {val2} ... ] that in an optimized manner pushes the given elements {val1} {val2}, etc. onto the front or back (i.e. l|r, left or right, head or tail) of the list tied to param {pm-name}.
  • Hash-set operation (i.e. hsh=( a b ...)) has been greately optimized for over-internet tied hash parameters.

• 2018-01-09

  • New option to ztie: -S, which used in conjunction with -L (lazy binding) causes database connection to be defered until first use of variable. Standard lazy binding means: key isn't required to exist.

• 2018-01-08

  • New option to ztie: -D, which causes mapped database key to be deleted on unset of the tied variable. Up to this moment this behavior was the default.

Redis can store strings at given keys, using SET command. Zredis maps those to hash array (like Zsh gdbm module):

% redis-cli -n 4 SET key1 value1
% redis-cli -n 4 SET key2 value2
% ztie -d db/redis -a "127.0.0.1/4" redis
% echo $zredis_tied
redis
% echo ${(kv)redis}
key1 value1 key2 value2

By appending /NAME to the host-spec (-f option), one can select single key of type HASH and map it to Zsh hash:

% redis-cli -n 4 hmset HASH key1 value1 key2 value2
% ztie -d db/redis -a "127.0.0.1/4/HASH" hset
% echo $zredis_tied
hset
% echo ${(kv)hset}
key1 value1 key2 value2
% echo $hset[key2]
value2
% unset 'hset[key2]'
% echo ${(kv)hset}
key1 value1

Can clear single elements by assigning () to array element. Can overwrite whole set by assigning via =( ... ) to set, and delete set from database by use of unset. Use zuntie to only detach variable from database without deleting any data.

% redis-cli -n 4 sadd SET value1 value2 value3 ''
% ztie -d db/redis -a "127.0.0.1/4/SET" myset
% echo ${myset[@]}
value2 value3 value1
% echo -E ${(qq)myset[@]}  # (qq) – quote with '', use to see empty elements
'value2' 'value3' '' 'value1'
% myset=( 1 2 3 )
% myset[2]=()
% redis-cli -n 4 smembers SET
1) "1"
3) "3"
% unset myset
% redis-cli -n 4 smembers SET
(empty list or set)

This variant maps zset as hash - keys are set elements, values are ranks. zrzset call outputs elements sorted according to the rank:

% redis-cli -n 4 zadd NEWZSET 1.0 a
% ztie -d db/redis -a "127.0.0.1/4/NEWZSET" zset
% echo ${(kv)zset}
a 1
% zset[a]=2.5
% zset[b]=1.5
% zrzset zset
% echo $reply
b a

There is no analogue of zrzset call because Zsh array already has correct order:

% redis-cli -n 4 rpush LIST value1 value2 value3
% ztie -d db/redis -a "127.0.0.1/4/LIST" mylist
% echo $mylist
value1 value2 value3
% mylist=( 1 2 3 )
% mylist[2]=()
% redis-cli -n 4 lrange LIST 0 -1
1) "1"
3) "3"
% zuntie mylist
% redis-cli -n 4 lrange LIST 0 -1
1) "1"
3) "3"

Single keys in main Redis storage are bound to Zsh string variables:

% redis-cli -n 4 KEYS "*"
1) "key1"
2) "SET"
3) "LIST"
4) "HASH"
5) "NEWZSET"
6) "key2"
% ztie -d db/redis -a "127.0.0.1/4/key1" key1
% echo $key1
value1
% key1=value2
% echo $key1
value2
% redis-cli -n 4 get key1
"value2"

The plugin is "standalone", which means that only sourcing it is needed. So to install, unpack zredis somewhere and add

source {where-zredis-is}/zredis.plugin.zsh

to zshrc.

If using a plugin manager, then Zplugin is recommended, but you can use any other too, and also install with Oh My Zsh (by copying directory to ~/.oh-my-zsh/custom/plugins).

Add zplugin light zdharma/zredis to your .zshrc file. Zplugin will handle cloning the plugin for you automatically the next time you start zsh. To update issue zplugin update zdharma/zredis.

Add antigen bundle zdharma/zredis to your .zshrc file. Antigen will handle cloning the plugin for you automatically the next time you start zsh.

1. cd ~/.oh-my-zsh/custom/plugins
2. git clone git@github.com:zdharma/zredis.git
3. Add zredis to your plugin list

Add zgen load zdharma/zredis to your .zshrc file in the same place you're doing your other zgen load calls in.

The values being set are the defaults. Change the values before loading zredis plugin.

zstyle ":plugin:zredis" cppflags "-I/usr/local/include"  # Additional include directory
zstyle ":plugin:zredis" cflags "-Wall -O2 -g"            # Additional CFLAGS
zstyle ":plugin:zredis" ldflags "-L/usr/local/lib"       # Additional library directory
zstyle ":plugin:zredis" configure_opts ""                # Additional ./configure options