This commit is contained in:
Thomas Gideon 2023-07-15 14:18:27 -04:00
parent e9f6051f60
commit b8a1c5461b
3 changed files with 286 additions and 108 deletions

148
Cargo.lock generated
View file

@ -41,6 +41,55 @@ dependencies = [
"libc",
]
[[package]]
name = "anstream"
version = "0.3.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0ca84f3628370c59db74ee214b3263d58f9aadd9b4fe7e711fd87dc452b7f163"
dependencies = [
"anstyle",
"anstyle-parse",
"anstyle-query",
"anstyle-wincon",
"colorchoice",
"is-terminal",
"utf8parse",
]
[[package]]
name = "anstyle"
version = "1.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3a30da5c5f2d5e72842e00bcb57657162cdabef0931f40e2deb9b4140440cecd"
[[package]]
name = "anstyle-parse"
version = "0.2.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "938874ff5980b03a87c5524b3ae5b59cf99b1d6bc836848df7bc5ada9643c333"
dependencies = [
"utf8parse",
]
[[package]]
name = "anstyle-query"
version = "1.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "5ca11d4be1bab0c8bc8734a9aa7bf4ee8316d462a08c6ac5052f888fef5b494b"
dependencies = [
"windows-sys",
]
[[package]]
name = "anstyle-wincon"
version = "1.0.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "180abfa45703aebe0093f79badacc01b8fd4ea2e35118747e5811127f926e188"
dependencies = [
"anstyle",
"windows-sys",
]
[[package]]
name = "anyhow"
version = "1.0.71"
@ -177,6 +226,53 @@ dependencies = [
"winapi",
]
[[package]]
name = "clap"
version = "4.3.12"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3eab9e8ceb9afdade1ab3f0fd8dbce5b1b2f468ad653baf10e771781b2b67b73"
dependencies = [
"clap_builder",
"clap_derive",
"once_cell",
]
[[package]]
name = "clap_builder"
version = "4.3.12"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9f2763db829349bf00cfc06251268865ed4363b93a943174f638daf3ecdba2cd"
dependencies = [
"anstream",
"anstyle",
"clap_lex",
"strsim",
]
[[package]]
name = "clap_derive"
version = "4.3.12"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "54a9bb5758fc5dfe728d1019941681eccaf0cf8a4189b692a0ee2f2ecf90a050"
dependencies = [
"heck",
"proc-macro2",
"quote",
"syn 2.0.25",
]
[[package]]
name = "clap_lex"
version = "0.5.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "2da6da31387c7e4ef160ffab6d5e7f00c42626fe39aea70a7b0f1773f7dd6c1b"
[[package]]
name = "colorchoice"
version = "1.0.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "acbf1af155f9b9ef647e42cdc158db4b64a1b61f743629225fde6f3e0be2a7c7"
[[package]]
name = "combine"
version = "4.6.6"
@ -520,6 +616,12 @@ version = "0.12.3"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "8a9ee70c43aaf417c914396645a0fa852624801b24ebb7ae78fe8272889ac888"
[[package]]
name = "heck"
version = "0.4.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "95505c38b4572b2d910cecb0281560f54b440a19336cbbcb27bf6ce6adc6f5a8"
[[package]]
name = "hermit-abi"
version = "0.3.2"
@ -772,6 +874,22 @@ dependencies = [
"wasm-bindgen",
]
[[package]]
name = "kensho"
version = "0.1.0"
dependencies = [
"anyhow",
"chrono",
"clap",
"env_logger",
"html2md",
"log",
"megalodon",
"rss",
"tokio",
"tokio-stream",
]
[[package]]
name = "lazy_static"
version = "1.4.0"
@ -1737,6 +1855,17 @@ dependencies = [
"tokio",
]
[[package]]
name = "tokio-stream"
version = "0.1.14"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "397c988d37662c7dda6d2208364a706264bf3d6138b11d436cbac0ad38832842"
dependencies = [
"futures-core",
"pin-project-lite",
"tokio",
]
[[package]]
name = "tokio-tungstenite"
version = "0.19.0"
@ -1883,6 +2012,12 @@ version = "0.7.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "09cc8ee72d2a9becf2f2febe0205bbed8fc6615b7cb429ad062dc7b7ddd036a9"
[[package]]
name = "utf8parse"
version = "0.2.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "711b9620af191e0cdc7468a8d14e709c3dcdb115b36f838e601583af800a370a"
[[package]]
name = "vcpkg"
version = "0.2.15"
@ -2005,19 +2140,6 @@ dependencies = [
"web-sys",
]
[[package]]
name = "watershed"
version = "0.1.0"
dependencies = [
"anyhow",
"chrono",
"env_logger",
"html2md",
"megalodon",
"rss",
"tokio",
]
[[package]]
name = "web-sys"
version = "0.3.64"

View file

@ -1,5 +1,5 @@
[package]
name = "watershed"
name = "kensho"
version = "0.1.0"
edition = "2021"
@ -8,8 +8,11 @@ edition = "2021"
[dependencies]
anyhow = "1.0.71"
chrono = "0.4.26"
clap = { version = "4.3.12", features = ["default", "derive"] }
env_logger = "0.10.0"
html2md = "0.2.14"
log = "0.4.19"
megalodon = "0.8.3"
rss = "2.0.4"
tokio = { version = "1.28.2", features = ["default", "full"] }
tokio-stream = "0.1.14"

View file

@ -1,70 +1,118 @@
use anyhow::{format_err, Result};
use chrono::{DateTime, Local, NaiveDate, TimeZone, Utc};
use anyhow::{bail, format_err, Result};
use chrono::{DateTime, Local, LocalResult, NaiveDate, TimeZone, Utc};
use html2md::parse_html;
use log::{debug, trace};
use megalodon::{
entities::status::Status, generator, megalodon::GetLocalTimelineInputOptions,
response::Response,
entities::Status, generator, megalodon::GetLocalTimelineInputOptions, response::Response,
Megalodon,
};
use tokio_stream::{iter, StreamExt};
use std::env;
#[derive(Debug)]
struct Post {
id: String,
content: String,
created_at: DateTime<Utc>,
struct Range {
start: DateTime<Local>,
end: DateTime<Local>,
}
// TODO implement try_from that looks for descendants and adds them
impl From<&Status> for Post {
fn from(status: &Status) -> Self {
let Status {
id,
created_at,
content,
..
} = status;
let id = id.clone();
let created_at = created_at.clone();
let content = parse_html(&content);
Post {
id,
created_at,
content,
}
}
#[derive(Debug)]
struct Page<'a> {
oldest_id: Option<String>,
oldest: Option<&'a DateTime<Utc>>,
newest: Option<&'a DateTime<Utc>>,
}
#[tokio::main]
async fn main() -> Result<()> {
env::set_var("RUST_LOG", format!("{}=debug", module_path!()));
env_logger::init();
// TODO add clap and argument for date
let start = Local
.from_local_datetime(
&NaiveDate::from_ymd_opt(2023, 7, 1)
.ok_or_else(|| format_err!("Invalid date!"))?
.and_hms_opt(0, 0, 0)
.expect("Failed to construct time!"),
let day = try_create_range("2023-07-01")?;
debug!("Date {}", day.end.format("%Y-%m-%d"));
let client = create_client()?;
let mut last_id_on_page: Option<String> = None;
debug!("Fetching posts");
loop {
let json = fetch_page(&client, &last_id_on_page).await?;
let page = Page {
newest: json.first().map(|s| &s.created_at),
oldest_id: json.last().map(|s| s.id.clone()),
oldest: json.last().map(|s| &s.created_at),
};
trace!("Page bounds {:?}", page);
if last_id_on_page.is_some() && page_start_older_than(&page, &day) {
break;
}
if let Some(oldest_id) = page_newer_than(&page, &day) {
last_id_on_page.replace(oldest_id);
continue;
}
let json = json
.clone()
.into_iter()
.filter(|json| day.start <= json.created_at && json.created_at <= day.end)
.collect::<Vec<Status>>();
trace!("Filtered to {} post(s)", json.len());
let mut stream = iter(json);
while let Some(status) = stream.next().await {
println!(
"{}
> {}",
status.created_at.with_timezone(&Local).format("%H:%M"),
parse_html(&status.content)
);
let Response { json, .. } = client.get_status_context(status.id, None).await?;
let thread = json
.descendants
.into_iter()
.map(|status| {
format!(
">
> {}
>> {}",
status.created_at.with_timezone(&Local).format("%H:%M"),
parse_html(&status.content)
)
.unwrap();
})
.collect::<Vec<String>>()
.join("\n");
println!("{}", thread);
}
let end = Local
.from_local_datetime(
&NaiveDate::from_ymd_opt(2023, 7, 1)
.ok_or_else(|| format_err!("Invallid date!"))?
.and_hms_opt(23, 59, 59)
.expect("Failed to construct time!"),
)
.unwrap();
if page_end_older_than(&page, &day) {
debug!("No more posts in range.");
break;
}
println!("Date {:#?}", start);
if let Some(id) = page.oldest_id {
last_id_on_page.replace(id.clone());
}
}
Ok(())
}
fn create_client() -> Result<Box<dyn Megalodon + Send + Sync>> {
let url = env::var("MASTODON_URL")?;
let token = env::var("MASTODON_ACCESS_TOKEN")?;
let client = generator(megalodon::SNS::Mastodon, url, Some(token), None);
let mut max_id: Option<String> = None;
loop {
let Response { json, .. } = if let Some(max_id) = max_id.as_ref() {
Ok(generator(megalodon::SNS::Mastodon, url, Some(token), None))
}
async fn fetch_page(
client: &Box<dyn Megalodon + Send + Sync>,
last_id_on_page: &Option<String>,
) -> Result<Vec<Status>> {
let Response { json, .. } = if let Some(max_id) = last_id_on_page.as_ref() {
debug!("Fetching next page");
client
.get_local_timeline(Some(&GetLocalTimelineInputOptions {
max_id: Some(max_id.clone()),
@ -72,56 +120,61 @@ async fn main() -> Result<()> {
}))
.await?
} else {
debug!("Fetching first page");
client.get_local_timeline(None).await?
};
Ok(json)
}
if let Some(last) = json.last() {
if last.created_at > start {
max_id.replace(last.id.clone());
continue;
}
}
println!(
"{}",
json.iter()
.filter(|json| start <= json.created_at && json.created_at <= end)
.map(Post::from)
.map(|post| {
format!(
"{} ({})
> {}",
post.created_at.format("%H:%M"),
post.id,
post.content
)
fn try_create_range<S: AsRef<str>>(date: S) -> Result<Range> {
Ok(Range {
start: create_day_bound(&date, 0, 0, 0)?,
end: create_day_bound(date, 23, 59, 59)?,
})
.collect::<Vec<String>>()
.join("\n\n")
);
let context = client
.get_status_context(String::from("110638913257555200"), None)
.await?;
println!(
"{}",
context
.json
.descendants
.iter()
.map(Post::from)
.map(|post| {
format!(
">
> {} ({})
>> {}",
post.created_at.format("%H:%M"),
post.id,
post.content
)
})
.collect::<Vec<String>>()
.join("\n")
);
break;
}
Ok(())
fn create_day_bound<S: AsRef<str>>(
day: S,
hour: u32,
minute: u32,
second: u32,
) -> Result<DateTime<Local>> {
let ts: Vec<&str> = day.as_ref().split("-").collect();
if ts.len() != 3 {
bail!("Invalid date format! {}", day.as_ref())
}
let (year, month, day) = if let [year, month, day, ..] = &ts[..] {
(year, month, day)
} else {
bail!("Invalid date format! {}", day.as_ref())
};
let b = Local.from_local_datetime(
&NaiveDate::from_ymd_opt(year.parse()?, month.parse()?, day.parse()?)
.ok_or_else(|| format_err!("Invalid date!"))?
.and_hms_opt(hour, minute, second)
.ok_or_else(|| format_err!("Invalid time!"))?,
);
if let LocalResult::Single(b) = b {
Ok(b)
} else {
bail!("Cannot construct day boundary!")
}
}
fn page_newer_than(page: &Page, range: &Range) -> Option<String> {
page.oldest
.filter(|oldest| *oldest > &range.end)
.and_then(|_| page.oldest_id.clone())
}
fn page_end_older_than(page: &Page, range: &Range) -> bool {
status_older_than(&page.oldest, &range.start)
}
fn page_start_older_than(page: &Page, range: &Range) -> bool {
status_older_than(&page.newest, &range.start)
}
fn status_older_than(status: &Option<&DateTime<Utc>>, dt: &DateTime<Local>) -> bool {
status.map(|status| status < dt).unwrap_or_default()
}