refactored model into future streams of streams idk izzy told me to do it

This commit is contained in:
Emile 2021-07-05 01:13:20 +01:00
parent 9b89311c44
commit f513f1d591
6 changed files with 223 additions and 84 deletions

133
Cargo.lock generated
View File

@ -66,6 +66,22 @@ dependencies = [
"futures-lite",
]
[[package]]
name = "async-global-executor"
version = "2.0.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9586ec52317f36de58453159d48351bc244bc24ced3effc1fce22f3d48664af6"
dependencies = [
"async-channel",
"async-executor",
"async-io",
"async-mutex",
"blocking",
"futures-lite",
"num_cpus",
"once_cell",
]
[[package]]
name = "async-io"
version = "1.6.0"
@ -94,6 +110,15 @@ dependencies = [
"event-listener",
]
[[package]]
name = "async-mutex"
version = "1.4.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "479db852db25d9dbf6204e6cb6253698f175c15726470f78af0d918e99d6156e"
dependencies = [
"event-listener",
]
[[package]]
name = "async-net"
version = "1.6.1"
@ -122,6 +147,34 @@ dependencies = [
"winapi 0.3.9",
]
[[package]]
name = "async-std"
version = "1.9.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d9f06685bad74e0570f5213741bea82158279a4103d988e57bfada11ad230341"
dependencies = [
"async-channel",
"async-global-executor",
"async-io",
"async-lock",
"async-process",
"crossbeam-utils 0.8.5",
"futures-channel",
"futures-core",
"futures-io",
"futures-lite",
"gloo-timers",
"kv-log-macro",
"log 0.4.14",
"memchr",
"num_cpus",
"once_cell",
"pin-project-lite",
"pin-utils",
"slab",
"wasm-bindgen-futures",
]
[[package]]
name = "async-task"
version = "4.0.3"
@ -440,7 +493,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9f02af974daeee82218205558e51ec8768b48cf524bd01d550abe5573a608285"
dependencies = [
"crossbeam-epoch",
"crossbeam-utils",
"crossbeam-utils 0.7.2",
"maybe-uninit",
]
@ -452,7 +505,7 @@ checksum = "058ed274caafc1f60c4997b5fc07bf7dc7cca454af7c6e81edffe5f33f70dace"
dependencies = [
"autocfg 1.0.1",
"cfg-if 0.1.10",
"crossbeam-utils",
"crossbeam-utils 0.7.2",
"lazy_static",
"maybe-uninit",
"memoffset",
@ -466,7 +519,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "774ba60a54c213d409d5353bda12d49cd68d14e45036a285234c8d6f91f92570"
dependencies = [
"cfg-if 0.1.10",
"crossbeam-utils",
"crossbeam-utils 0.7.2",
"maybe-uninit",
]
@ -481,6 +534,26 @@ dependencies = [
"lazy_static",
]
[[package]]
name = "crossbeam-utils"
version = "0.8.5"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "d82cfc11ce7f2c3faef78d8a684447b40d503d9681acebed6cb728d45940c4db"
dependencies = [
"cfg-if 1.0.0",
"lazy_static",
]
[[package]]
name = "ctor"
version = "0.1.20"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5e98e2ad1a782e33928b96fc3948e7c355e5af34ba4de7670fe8bac2a3b2006d"
dependencies = [
"quote",
"syn",
]
[[package]]
name = "data-encoding"
version = "2.3.2"
@ -853,6 +926,19 @@ version = "0.3.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9b919933a397b79c37e33b77bb2aa3dc8eb6e165ad809e58ff75bc7db2e34574"
[[package]]
name = "gloo-timers"
version = "0.2.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "47204a46aaff920a1ea58b11d03dec6f704287d27561724a4631e450654a891f"
dependencies = [
"futures-channel",
"futures-core",
"js-sys",
"wasm-bindgen",
"web-sys",
]
[[package]]
name = "groupable"
version = "0.2.0"
@ -1107,6 +1193,7 @@ checksum = "dd25036021b0de88a0aff6b850051563c6516d0bf53f8638938edbb9de732736"
name = "izzilis"
version = "0.1.0"
dependencies = [
"async-std",
"chrono",
"elefren",
"frankenstein",
@ -1137,6 +1224,15 @@ dependencies = [
"winapi-build",
]
[[package]]
name = "kv-log-macro"
version = "1.0.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0de8b303297635ad57c9f5059fd9cee7a47f8e8daa09df0fcd07dd39fb22977f"
dependencies = [
"log 0.4.14",
]
[[package]]
name = "language-tags"
version = "0.2.2"
@ -1180,6 +1276,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "51b9bbe6c47d51fc3e1a9b945965946b4c44142ab8792c50835a980d362c2710"
dependencies = [
"cfg-if 1.0.0",
"value-bag",
]
[[package]]
@ -2375,7 +2472,7 @@ version = "0.1.10"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "fb2d1b8f4548dbf5e1f7818512e9c406860678f29c300cdf0ebac72d1a3a1671"
dependencies = [
"crossbeam-utils",
"crossbeam-utils 0.7.2",
"futures 0.1.31",
]
@ -2396,7 +2493,7 @@ version = "0.1.12"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "09bc590ec4ba8ba87652da2068d150dcada2cfa2e07faae270a5e0409aa51351"
dependencies = [
"crossbeam-utils",
"crossbeam-utils 0.7.2",
"futures 0.1.31",
"lazy_static",
"log 0.4.14",
@ -2441,7 +2538,7 @@ checksum = "df720b6581784c118f0eb4310796b12b1d242a7eb95f716a8367855325c25f89"
dependencies = [
"crossbeam-deque",
"crossbeam-queue",
"crossbeam-utils",
"crossbeam-utils 0.7.2",
"futures 0.1.31",
"lazy_static",
"log 0.4.14",
@ -2456,7 +2553,7 @@ version = "0.2.13"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "93044f2d313c95ff1cb7809ce9a7a05735b012288a888b62d4434fd58c94f296"
dependencies = [
"crossbeam-utils",
"crossbeam-utils 0.7.2",
"futures 0.1.31",
"slab",
"tokio-executor",
@ -2659,6 +2756,16 @@ dependencies = [
"rand 0.6.5",
]
[[package]]
name = "value-bag"
version = "1.0.0-alpha.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "dd320e1520f94261153e96f7534476ad869c14022aee1e59af7c778075d840ae"
dependencies = [
"ctor",
"version_check 0.9.3",
]
[[package]]
name = "vcpkg"
version = "0.2.15"
@ -2742,6 +2849,18 @@ dependencies = [
"wasm-bindgen-shared",
]
[[package]]
name = "wasm-bindgen-futures"
version = "0.4.24"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5fba7978c679d53ce2d0ac80c8c175840feb849a161664365d1287b41f2e67f1"
dependencies = [
"cfg-if 1.0.0",
"js-sys",
"wasm-bindgen",
"web-sys",
]
[[package]]
name = "wasm-bindgen-macro"
version = "0.2.74"

View File

@ -7,6 +7,7 @@ edition = "2018"
# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
[dependencies]
async-std = { version = "1.9.0", features = ["unstable"] }
chrono = "0.4.19"
elefren = { version = "0.22.0", features = ["toml"] }
frankenstein = "0.4.0"

View File

@ -1,11 +1,14 @@
use rand::Rng;
use std::error::Error;
use crate::{generator, model, publish, selection};
use crate::{
model::{self, SampleModel},
publish, selection,
};
pub struct IzzilisBot<T: model::SampleModel, U: publish::Publisher, V: selection::Selector> {
generator: generator::Generator<T>,
publisher: U, // One day I'll figure out how to make this a vector with differing Publisher types
pub struct IzzilisBot<T: SampleModel, U: publish::Publisher, V: selection::Selector> {
model: T,
publisher: U,
selector: V,
loaded_samples: Vec<String>,
}
@ -16,13 +19,9 @@ where
U: publish::Publisher,
V: selection::Selector,
{
pub fn new(
generator: generator::Generator<T>,
publisher: U,
selector: V,
) -> IzzilisBot<T, U, V> {
pub fn new(model: T, publisher: U, selector: V) -> IzzilisBot<T, U, V> {
Self {
generator,
model,
publisher,
loaded_samples: Vec::new(),
selector: selector,
@ -30,18 +29,19 @@ where
}
pub fn generate_samples(&mut self) -> Result<(), Box<dyn Error>> {
let lines = self.generator.generate_sample_lines()?;
for line in lines {
self.selector.send_for_review(line)?;
}
// let lines = self.model.generate_sample_lines()?;
// for line in lines {
// self.selector.send_for_review(line)?;
// }
self.loaded_samples = self.selector.collect_selected_samples(); // wtf happens to the original self.loaded_samples???????
// self.loaded_samples = self.selector.collect_selected_samples(); // wtf happens to the original self.loaded_samples???????
Ok(())
// Ok(())
todo!()
}
pub fn publish(&mut self) -> Result<(), U::Error> {
if self.loaded_samples.len() == 0 {
if self.loaded_samples.len() < 5 {
// Refresh samples. Either none have been generated so far,
// or generated ones are stale.
//

View File

@ -1,30 +0,0 @@
use std::io;
use crate::model;
const SAMPLE_SPLIT_WORD: &str = "<|endoftext|>";
const SAMPLE_SAMPLE_LINE: &str =
"======================================== SAMPLE 1 ========================================";
pub struct Generator<T: model::SampleModel> {
model: T,
}
// Why did this fucking shit take so long to sort out??
impl<T> Generator<T>
where
T: model::SampleModel,
{
pub fn new(model: T) -> Generator<T> {
Self { model }
}
pub fn generate_sample_lines(&self) -> Result<Vec<String>, io::Error> {
Ok(self
.model
.get_sample()?
.replace(SAMPLE_SAMPLE_LINE, "")
.split(SAMPLE_SPLIT_WORD)
.into_iter()
.map(|elem| elem.trim().to_owned())
.collect::<Vec<_>>())
}
}

View File

@ -3,15 +3,14 @@ use std::{error::Error, process, time::Duration};
use chrono::Local;
use rand::Rng;
use crate::{
bot::IzzilisBot, generator::Generator, publish::FediversePublisher, selection::ConsoleSelector,
};
use crate::{bot::IzzilisBot, publish::FediversePublisher, selection::ConsoleSelector};
use futures::StreamExt;
use futures_timer::Delay;
use model::SampleModelExt;
mod bot;
mod config;
mod generator;
mod model;
mod publish;
mod selection;
@ -35,7 +34,7 @@ fn main() -> Result<(), Box<dyn Error>> {
}
};
let gpt_model = model::GPTSampleModel::new(
let mut gpt_model = model::GPTSampleModel::new(
cfg.python_path(),
cfg.gpt_code_path(),
vec![
@ -49,25 +48,32 @@ fn main() -> Result<(), Box<dyn Error>> {
"--nsamples".to_string(),
"1".to_string(),
],
);
let publisher = FediversePublisher::new(cfg.fediverse_base_url())?;
// let publisher = ConsolePublisher::new();
let gen = Generator::new(gpt_model);
let console_selector = ConsoleSelector::new();
let mut bot = IzzilisBot::new(gen, publisher, console_selector);
bot.generate_samples();
)
.into_stream()
.take(60);
let cfg_interval = cfg.interval_seconds();
loop {
let wait_seconds = rand::thread_rng().gen_range(cfg_interval.min()..cfg_interval.max());
let wait_time = Duration::from_secs(wait_seconds);
let now = Local::now();
println!("[{}] Next post is in [{}] seconds", now, wait_seconds);
Delay::new(wait_time).await;
match bot.publish() {
Err(err) => println!("Got error from publish: [{}]; continuing", err),
Ok(()) => println!("publish() call successful"),
}
while let Some(Ok(sample)) = gpt_model.next().await {
println!("{}", sample);
}
return Ok(());
// let publisher = FediversePublisher::new(cfg.fediverse_base_url())?;
// // let publisher = ConsolePublisher::new();
// let console_selector = ConsoleSelector::new();
// let mut bot = IzzilisBot::new(gen, publisher, console_selector);
// bot.generate_samples();
// let cfg_interval = cfg.interval_seconds();
// loop {
// let wait_seconds = rand::thread_rng().gen_range(cfg_interval.min()..cfg_interval.max());
// let wait_time = Duration::from_secs(wait_seconds);
// let now = Local::now();
// println!("[{}] Next post is in [{}] seconds", now, wait_seconds);
// Delay::new(wait_time).await;
// match bot.publish() {
// Err(err) => println!("Got error from publish: [{}]; continuing", err),
// Ok(()) => println!("publish() call successful"),
// }
// }
})
}

View File

@ -1,7 +1,11 @@
use std::{io, process::Command};
use async_std::{io, process::Command};
use futures::{future::BoxFuture, stream::BoxStream, Future, Stream, StreamExt, TryStreamExt};
pub trait SampleModel {
fn get_sample(&self) -> Result<String, io::Error>;
type Error;
type Sample: Future<Output = Result<String, Self::Error>>;
fn get_sample(&self) -> Self::Sample;
}
pub struct GPTSampleModel {
@ -11,13 +15,15 @@ pub struct GPTSampleModel {
}
impl SampleModel for GPTSampleModel {
fn get_sample(&self) -> Result<String, io::Error> {
let cmd_output = Command::new(&self.python_command)
type Error = io::Error;
type Sample = BoxFuture<'static, Result<String, Self::Error>>;
fn get_sample(&self) -> Self::Sample {
let cmd = Command::new(&self.python_command)
.current_dir(&self.command_working_path)
.args(&self.command_args)
.output()?;
Ok(String::from_utf8_lossy(&cmd_output.stdout).to_string())
.output();
Box::pin(async { Ok(String::from_utf8_lossy(&cmd.await?.stdout).to_string()) })
}
}
@ -34,3 +40,40 @@ impl GPTSampleModel {
}
}
}
pub trait SampleModelExt: SampleModel {
type Stream: Stream<Item = Result<String, Self::Error>>;
fn into_stream(self) -> Self::Stream;
}
const SAMPLE_SPLIT_WORD: &str = "<|endoftext|>";
const SAMPLE_SAMPLE_LINE: &str =
"======================================== SAMPLE 1 ========================================";
impl<T: SampleModel + Send + Sync + 'static> SampleModelExt for T
where
Self::Sample: Send,
{
type Stream = BoxStream<'static, Result<String, Self::Error>>;
fn into_stream(self) -> Self::Stream {
Box::pin(
futures::stream::try_unfold(self, |this| async {
Ok(Some((this.get_sample().await?, this)))
})
.map_ok(|samples| {
futures::stream::iter(
samples
.replace(SAMPLE_SAMPLE_LINE, "")
.split(SAMPLE_SPLIT_WORD)
.map(|elem| elem.to_owned())
.collect::<Vec<String>>()
.into_iter()
.map(|elem| Ok(elem.trim().to_owned())),
)
})
.try_flatten(),
)
}
}