Add support for yaml datetimes (#2208)

* Add support for yaml datetimes

* cargo fmt

* Wire things up; Add more tests

* cargo fmt

Cargo.lock

@@ -4632,6 +4632,7 @@ dependencies = [
  "libs",
  "serde",
  "tempfile",
+ "time",
 ]
 
 [[package]]

components/content/src/front_matter/page.rs

@@ -7,7 +7,7 @@ use time::macros::{format_description, time};
 use time::{Date, OffsetDateTime, PrimitiveDateTime};
 
 use errors::{bail, Result};
-use utils::de::{fix_toml_dates, from_toml_datetime};
+use utils::de::{fix_toml_dates, from_unknown_datetime};
 
 use crate::front_matter::split::RawFrontMatter;
 
@@ -20,7 +20,7 @@ pub struct PageFrontMatter {
     /// Description in <meta> that appears when linked, e.g. on twitter
     pub description: Option<String>,
     /// Updated date
-    #[serde(default, deserialize_with = "from_toml_datetime")]
+    #[serde(default, deserialize_with = "from_unknown_datetime")]
     pub updated: Option<String>,
     /// Datetime content was last updated
     #[serde(default, skip_deserializing)]
@@ -29,7 +29,7 @@ pub struct PageFrontMatter {
     #[serde(default, skip_deserializing)]
     pub updated_datetime_tuple: Option<(i32, u8, u8)>,
     /// Date if we want to order pages (ie blog post)
-    #[serde(default, deserialize_with = "from_toml_datetime")]
+    #[serde(default, deserialize_with = "from_unknown_datetime")]
     pub date: Option<String>,
     /// Datetime content was created
     #[serde(default, skip_deserializing)]
@@ -129,6 +129,7 @@ impl PageFrontMatter {
     /// Converts the TOML datetime to a time::OffsetDateTime
     /// Also grabs the year/month/day tuple that will be used in serialization
     pub fn date_to_datetime(&mut self) {
+        println!("{:?}", self.date);
         self.datetime = self.date.as_ref().map(|s| s.as_ref()).and_then(parse_datetime);
         self.datetime_tuple = self.datetime.map(|dt| (dt.year(), dt.month().into(), dt.day()));
 
@@ -333,6 +334,36 @@ date: 2002-10-02T15:00:00.123456Z
         assert_eq!(res.datetime.unwrap(), datetime!(2002 - 10 - 02 15:00:00.123456 UTC));
     }
 
+    #[test_case(&RawFrontMatter::Yaml(r#"
+title: Hello
+description: hey there
+date: 2001-12-15T02:59:43.1Z
+"#); "canonical")]
+    #[test_case(&RawFrontMatter::Yaml(r#"
+title: Hello
+description: hey there
+date: 2001-12-14t21:59:43.10-05:00
+"#); "iso8601")]
+    #[test_case(&RawFrontMatter::Yaml(r#"
+title: Hello
+description: hey there
+date: 2001-12-14 21:59:43.10 -5
+"#); "space separated")]
+    #[test_case(&RawFrontMatter::Yaml(r#"
+title: Hello
+description: hey there
+date: 2001-12-15 2:59:43.10
+"#); "no time zone")]
+    #[test_case(&RawFrontMatter::Yaml(r#"
+title: Hello
+description: hey there
+date: 2001-12-15
+"#); "date only")]
+    fn can_parse_yaml_dates(content: &RawFrontMatter) {
+        let res = PageFrontMatter::parse(content).unwrap();
+        assert!(res.datetime.is_some());
+    }
+
     #[test_case(&RawFrontMatter::Toml(r#"
 title = "Hello"
 description = "hey there"

components/utils/Cargo.toml

@@ -13,3 +13,4 @@ libs = { path = "../libs" }
 
 [dev-dependencies]
 tempfile = "3"
+time = { version = "0.3", features = ["macros"] }

components/utils/src/de.rs

@@ -1,12 +1,83 @@
+use core::convert::TryFrom;
+use errors::{anyhow, Result};
+use libs::regex::Regex;
 use libs::tera::{Map, Value};
+use libs::time;
+use libs::time::format_description::well_known::Rfc3339;
 use libs::toml;
 use serde::{Deserialize, Deserializer};
 
+pub fn parse_yaml_datetime(date_string: &str) -> Result<time::OffsetDateTime> {
+    // See https://github.com/getzola/zola/issues/2071#issuecomment-1530610650
+    let re = Regex::new(r#"^"?([0-9]{4})-([0-9][0-9]?)-([0-9][0-9]?)([Tt]|[ \t]+)([0-9][0-9]?):([0-9]{2}):([0-9]{2})\.([0-9]*)?Z?([ \t]([-+][0-9][0-9]?)(:([0-9][0-9]?))?Z?|([-+][0-9]{2})?:([0-9]{2})?)?|([0-9]{4})-([0-9]{2})-([0-9]{2})"?$"#).unwrap();
+    let captures = if let Some(captures_) = re.captures(date_string) {
+        Ok(captures_)
+    } else {
+        Err(anyhow!("Error parsing YAML datetime"))
+    }?;
+    let year =
+        if let Some(cap) = captures.get(1) { cap } else { captures.get(15).unwrap() }.as_str();
+    let month =
+        if let Some(cap) = captures.get(2) { cap } else { captures.get(16).unwrap() }.as_str();
+    let day =
+        if let Some(cap) = captures.get(3) { cap } else { captures.get(17).unwrap() }.as_str();
+    let hours = if let Some(hours_) = captures.get(5) { hours_.as_str() } else { "0" };
+    let minutes = if let Some(minutes_) = captures.get(6) { minutes_.as_str() } else { "0" };
+    let seconds = if let Some(seconds_) = captures.get(7) { seconds_.as_str() } else { "0" };
+    let fractional_seconds_raw =
+        if let Some(fractionals) = captures.get(8) { fractionals.as_str() } else { "" };
+    let fractional_seconds_intermediate = fractional_seconds_raw.trim_end_matches("0");
+    //
+    // Prepare for eventual conversion into nanoseconds
+    let fractional_seconds = if fractional_seconds_intermediate.len() > 0
+        && fractional_seconds_intermediate.len() <= 9
+    {
+        fractional_seconds_intermediate
+    } else {
+        "0"
+    };
+    let maybe_timezone_hour_1 = captures.get(10);
+    let maybe_timezone_minute_1 = captures.get(12);
+    let maybe_timezone_hour_2 = captures.get(13);
+    let maybe_timezone_minute_2 = captures.get(14);
+    let maybe_timezone_hour;
+    let maybe_timezone_minute;
+    if maybe_timezone_hour_2.is_some() {
+        maybe_timezone_hour = maybe_timezone_hour_2;
+        maybe_timezone_minute = maybe_timezone_minute_2;
+    } else {
+        maybe_timezone_hour = maybe_timezone_hour_1;
+        maybe_timezone_minute = maybe_timezone_minute_1;
+    }
+
+    let mut offset_datetime = time::OffsetDateTime::UNIX_EPOCH;
+
+    if let Some(hour) = maybe_timezone_hour {
+        let minute_str =
+            if let Some(minute_) = maybe_timezone_minute { minute_.as_str() } else { "0" };
+        offset_datetime = offset_datetime.to_offset(time::UtcOffset::from_hms(
+            hour.as_str().parse()?,
+            minute_str.parse()?,
+            0,
+        )?);
+    }
+
+    // Free parse unwraps since we know everything is a digit courtesy of prior regex.
+    Ok(offset_datetime
+        .replace_year(year.parse().unwrap())?
+        .replace_month(time::Month::try_from(month.parse::<u8>().unwrap())?)?
+        .replace_day(day.parse().unwrap())?
+        .replace_hour(hours.parse().unwrap())?
+        .replace_minute(minutes.parse().unwrap())?
+        .replace_second(seconds.parse().unwrap())?
+        .replace_nanosecond(fractional_seconds.parse::<u32>().unwrap() * 100_000_000)?)
+}
+
 /// Used as an attribute when we want to convert from TOML to a string date
 /// If a TOML datetime isn't present, it will accept a string and push it through
 /// TOML's date time parser to ensure only valid dates are accepted.
 /// Inspired by this proposal: <https://github.com/alexcrichton/toml-rs/issues/269>
-pub fn from_toml_datetime<'de, D>(deserializer: D) -> Result<Option<String>, D::Error>
+pub fn from_unknown_datetime<'de, D>(deserializer: D) -> Result<Option<String>, D::Error>
 where
     D: Deserializer<'de>,
 {
@@ -22,10 +93,19 @@ where
 
     match MaybeDatetime::deserialize(deserializer)? {
         MaybeDatetime::Datetime(d) => Ok(Some(d.to_string())),
-        MaybeDatetime::String(s) => match toml::value::Datetime::from_str(&s) {
+        MaybeDatetime::String(s) => {
-            Ok(d) => Ok(Some(d.to_string())),
+            if let Ok(d) = toml::value::Datetime::from_str(&s) {
-            Err(e) => Err(D::Error::custom(e)),
+                Ok(Some(d.to_string()))
-        },
+            } else if let Ok(d) = parse_yaml_datetime(&s) {
+                // Ensure that the resulting string is easily reparseable down the line.
+                // In content::front_matter::page.rs where these strings are currently used,
+                // Rfc3339 works with the explicit demands in that code but not always with the result of
+                // _to_string.
+                Ok(Some(d.format(&Rfc3339).unwrap()))
+            } else {
+                Err(D::Error::custom("Unable to parse datetime"))
+            }
+        }
     }
 }
 
@@ -80,3 +160,62 @@ pub fn fix_toml_dates(table: Map<String, Value>) -> Value {
 
     Value::Object(new)
 }
+
+#[cfg(test)]
+mod tests {
+    use super::parse_yaml_datetime;
+    use time::macros::datetime;
+
+    #[test]
+    fn yaml_spec_examples_pass() {
+        let canonical = "2001-12-15T02:59:43.1Z";
+        let valid_iso8601 = "2001-12-14t21:59:43.10-05:00";
+        let space_separated = "2001-12-14 21:59:43.10 -5";
+        let no_time_zone = "2001-12-15 2:59:43.10";
+        let date = "2002-12-14";
+        assert_eq!(parse_yaml_datetime(canonical).unwrap(), datetime!(2001-12-15 2:59:43.1 +0));
+        assert_eq!(
+            parse_yaml_datetime(valid_iso8601).unwrap(),
+            datetime!(2001-12-14 21:59:43.1 -5)
+        );
+        assert_eq!(
+            parse_yaml_datetime(space_separated).unwrap(),
+            datetime!(2001-12-14 21:59:43.1 -5)
+        );
+        assert_eq!(parse_yaml_datetime(no_time_zone).unwrap(), datetime!(2001-12-15 2:59:43.1 +0));
+        assert_eq!(parse_yaml_datetime(date).unwrap(), datetime!(2002-12-14 0:00:00 +0));
+    }
+
+    #[test]
+    fn yaml_spec_invalid_dates_fail() {
+        let invalid_month = "2001-13-15";
+        assert!(parse_yaml_datetime(invalid_month).is_err());
+
+        let invalid_month = "2001-13-15T02:59:43.1Z";
+        assert!(parse_yaml_datetime(invalid_month).is_err());
+
+        let no_digits_in_year = "xxxx-12-15";
+        assert!(parse_yaml_datetime(no_digits_in_year).is_err());
+
+        let no_digits_in_year = "xxxx-12-15T02:59:43.1Z";
+        assert!(parse_yaml_datetime(no_digits_in_year).is_err());
+
+        let no_digits_in_month = "2001-xx-15";
+        assert!(parse_yaml_datetime(no_digits_in_month).is_err());
+
+        let no_digits_in_month = "2001-xx-15T02:59:43.1Z";
+        assert!(parse_yaml_datetime(no_digits_in_month).is_err());
+
+        let no_digits_in_day = "2001-12-xx";
+        assert!(parse_yaml_datetime(no_digits_in_day).is_err());
+
+        let no_digits_in_day = "2001-12-xx:59:43.1Z";
+        assert!(parse_yaml_datetime(no_digits_in_day).is_err());
+
+        let unparseable_time = "2001-12-15:69:43.1Z";
+        assert!(parse_yaml_datetime(unparseable_time).is_err());
+
+        let unparseable_time = "2001-12-15:59:4x.1Z";
+        assert!(parse_yaml_datetime(unparseable_time).is_err());
+    }
+}