If you need a simple Telegram bot capable of executing scripts (written in any language) and responding with text and pictures, then under cat. There you will find a story about a bot engine that does what you want.
A short list of engine features:
- The engine can serve several bots at once
- Bot executes scripts (written in any language)
- , ( )
- ( )
- HTTP, , "" ( cron-)
: , IP-, SSL-… , . , , .
, .
Go. , root-. , , , , .
:
cd tmp
git clone https://github.com/michurin/cnbot.git
cd cnbot
go build ./cmd/...
./cnbot
( ) , . , .
, . : Telegram.
(config.yaml
):
bots:
firstBot:
token: "22222222:AAAAAAAAAAAAAA"
script: "/usr/bin/true"
( - , /bin/rm
). (-i
):
./cnbot -i -c config.yaml
, .
( -i
)
./cnbot -c config.yaml
Telegram-.
user 500050880 is not allowed
user_id
( ),
bots:
firstBot:
token: "22222222:AAAAAAAAAAAAAA"
script: "/bin/echo"
allowed_users: [500050880]
, echo
script
. ( ) echo-. . hi
, Hi!
, -n hi
.
(-n
echo
). : ; , , , ; .
. /bin/echo
:
#!/bin/sh
env
Hello!
BOT_TEXT=Hello!
BOT_FROM_FIRSTNAME=Alexey
BOT_NAME=firstBot
BOT_CHAT=500050880
BOT_FROM=500050880
, , ID , , ( ; , ).
, . , white list . ID — . . demo.sh.
, PATH
. , .
#!/bin/sh
echo '%!PRE'
env | sort
. markdown, . demo.sh.
, stdout
:
#!/bin/sh
curl -qfs https://golang.org/lib/godoc/images/footer-gopher.jpg
, "empty", , "".
demo-, .
HTTP bind_address
:
bots:
firstBot:
token: "22222222:AAAAAAAAAAAAAA"
script: "/bin/echo"
allowed_users: [500050880]
bind_address: ":9091"
:
echo "ok" | curl -qfsX POST --data-binary @- "http://:9091/500050880"
, , . , output . , .
multipart/form-data
:
curl -qfsX POST -F to=500050880 -F msg=ok "http://:9091"
, . , , . , demo-. http- . readme , , , .
, , — .
!
...
..., , .
- . , . , , . , . .
, . : , , VPN, … , , - , .
, , , Telegram API . - bash+curl+jq, , .
. . , , - .
, . (), … , , . , , , .
2018 Go , , Go. :-) , , Go . , Go, Go, "" ( ).
. cnbot, .
?
, . . - , — . .
But I would like to develop the engine towards embeddability: to single out some simple part from it that could be connected by a library to any Go program. Conventionally, if you already have a microservice for ... what is it good for ... even for rendering 3D scenes, you can embed a bot chat into it in one or two lines for operational diagnostics / monitoring / control ... This, it seems to me, would be useful. In fact, you can do that now. It's just not very convenient.