Add from csv and from tsv (#320)

Cargo.lock

@@ -121,6 +121,7 @@ dependencies = [
  "lazy_static",
  "memchr",
  "regex-automata",
+ "serde",
 ]
 
 [[package]]
@@ -326,6 +327,28 @@ dependencies = [
  "winapi",
 ]
 
+[[package]]
+name = "csv"
+version = "1.1.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "22813a6dc45b335f9bade10bf7271dc477e81113e89eb251a0bc2a8a81c536e1"
+dependencies = [
+ "bstr",
+ "csv-core",
+ "itoa",
+ "ryu",
+ "serde",
+]
+
+[[package]]
+name = "csv-core"
+version = "0.1.10"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2b2466559f260f48ad25fe6317b3c8dac77b5bdb5763ac7d9d6103530663bc90"
+dependencies = [
+ "memchr",
+]
+
 [[package]]
 name = "ctor"
 version = "0.1.21"
@@ -719,6 +742,7 @@ dependencies = [
  "chrono",
  "chrono-humanize",
  "chrono-tz",
+ "csv",
  "dialoguer",
  "glob",
  "lscolors",

crates/nu-command/Cargo.toml

@@ -18,6 +18,7 @@ trash = { version = "1.3.0", optional = true }
 unicode-segmentation = "1.8.0"
 
 # Potential dependencies for extras
+csv = "1.1.3"
 glob = "0.3.0"
 Inflector = "0.11"
 thiserror = "1.0.29"

crates/nu-command/src/default_context.rs

@@ -45,7 +45,9 @@ pub fn create_default_context() -> EngineState {
             For,
             Format,
             From,
+            FromCsv,
             FromJson,
+            FromTsv,
             Get,
             Griddle,
             Help,

crates/nu-command/src/formats/from/csv.rs

@@ -0,0 +1,113 @@
+use super::delimited::from_delimited_data;
+
+use nu_engine::CallExt;
+use nu_protocol::ast::Call;
+use nu_protocol::engine::{Command, EngineState, Stack};
+use nu_protocol::{Example, PipelineData, ShellError, Signature, SyntaxShape, Value};
+
+#[derive(Clone)]
+pub struct FromCsv;
+
+impl Command for FromCsv {
+    fn name(&self) -> &str {
+        "from csv"
+    }
+
+    fn signature(&self) -> Signature {
+        Signature::build("from csv")
+            .named(
+                "separator",
+                SyntaxShape::String,
+                "a character to separate columns, defaults to ','",
+                Some('s'),
+            )
+            .switch(
+                "noheaders",
+                "don't treat the first row as column names",
+                Some('n'),
+            )
+    }
+
+    fn usage(&self) -> &str {
+        "Parse text as .csv and create table."
+    }
+
+    fn run(
+        &self,
+        engine_state: &EngineState,
+        stack: &mut Stack,
+        call: &Call,
+        input: PipelineData,
+    ) -> Result<nu_protocol::PipelineData, ShellError> {
+        from_csv(engine_state, stack, call, input)
+    }
+
+    fn examples(&self) -> Vec<Example> {
+        vec![
+            Example {
+                description: "Convert comma-separated data to a table",
+                example: "open data.txt | from csv",
+                result: None,
+            },
+            Example {
+                description: "Convert comma-separated data to a table, ignoring headers",
+                example: "open data.txt | from csv --noheaders",
+                result: None,
+            },
+            Example {
+                description: "Convert comma-separated data to a table, ignoring headers",
+                example: "open data.txt | from csv -n",
+                result: None,
+            },
+            Example {
+                description: "Convert semicolon-separated data to a table",
+                example: "open data.txt | from csv --separator ';'",
+                result: None,
+            },
+        ]
+    }
+}
+
+fn from_csv(
+    engine_state: &EngineState,
+    stack: &mut Stack,
+    call: &Call,
+    input: PipelineData,
+) -> Result<PipelineData, ShellError> {
+    let name = call.head;
+
+    let noheaders = call.has_flag("noheaders");
+    let separator: Option<Value> = call.get_flag(engine_state, stack, "separator")?;
+
+    let sep = match separator {
+        Some(Value::String { val: s, span }) => {
+            if s == r"\t" {
+                '\t'
+            } else {
+                let vec_s: Vec<char> = s.chars().collect();
+                if vec_s.len() != 1 {
+                    return Err(ShellError::MissingParameter(
+                        "single character separator".into(),
+                        span,
+                    ));
+                };
+                vec_s[0]
+            }
+        }
+        _ => ',',
+    };
+
+    from_delimited_data(noheaders, sep, input, name)
+}
+
+#[cfg(test)]
+mod test {
+    use super::*;
+
+    #[test]
+    fn test_examples() {
+        use crate::test_examples;
+
+        test_examples(FromCsv {})
+    }
+}

crates/nu-command/src/formats/from/delimited.rs

@@ -0,0 +1,61 @@
+use csv::ReaderBuilder;
+use nu_protocol::{IntoPipelineData, PipelineData, ShellError, Span, Value};
+
+fn from_delimited_string_to_value(
+    s: String,
+    noheaders: bool,
+    separator: char,
+    span: Span,
+) -> Result<Value, csv::Error> {
+    let mut reader = ReaderBuilder::new()
+        .has_headers(!noheaders)
+        .delimiter(separator as u8)
+        .from_reader(s.as_bytes());
+
+    let headers = if noheaders {
+        (1..=reader.headers()?.len())
+            .map(|i| format!("Column{}", i))
+            .collect::<Vec<String>>()
+    } else {
+        reader.headers()?.iter().map(String::from).collect()
+    };
+
+    let mut rows = vec![];
+    for row in reader.records() {
+        let mut output_row = vec![];
+        for value in row?.iter() {
+            if let Ok(i) = value.parse::<i64>() {
+                output_row.push(Value::Int { val: i, span });
+            } else if let Ok(f) = value.parse::<f64>() {
+                output_row.push(Value::Float { val: f, span });
+            } else {
+                output_row.push(Value::String {
+                    val: value.into(),
+                    span,
+                });
+            }
+        }
+        rows.push(Value::Record {
+            cols: headers.clone(),
+            vals: output_row,
+            span,
+        });
+    }
+
+    Ok(Value::List { vals: rows, span })
+}
+
+pub fn from_delimited_data(
+    noheaders: bool,
+    sep: char,
+    input: PipelineData,
+    name: Span,
+) -> Result<PipelineData, ShellError> {
+    let concat_string = input.collect_string("");
+
+    Ok(
+        from_delimited_string_to_value(concat_string, noheaders, sep, name)
+            .map_err(|x| ShellError::DelimiterError(x.to_string(), name))?
+            .into_pipeline_data(),
+    )
+}

crates/nu-command/src/formats/from/json.rs

@@ -77,7 +77,7 @@ impl Command for FromJson {
         input: PipelineData,
     ) -> Result<nu_protocol::PipelineData, ShellError> {
         let span = call.head;
-        let mut string_input = input.collect_string();
+        let mut string_input = input.collect_string("");
         string_input.push('\n');
 
         // TODO: turn this into a structured underline of the nu_json error

crates/nu-command/src/formats/from/mod.rs

@@ -1,5 +1,10 @@
 mod command;
+mod csv;
+mod delimited;
 mod json;
+mod tsv;
 
+pub use self::csv::FromCsv;
 pub use command::From;
 pub use json::FromJson;
+pub use tsv::FromTsv;

crates/nu-command/src/formats/from/tsv.rs

@@ -0,0 +1,56 @@
+use super::delimited::from_delimited_data;
+
+use nu_protocol::ast::Call;
+use nu_protocol::engine::{Command, EngineState, Stack};
+use nu_protocol::{PipelineData, ShellError, Signature};
+
+#[derive(Clone)]
+pub struct FromTsv;
+
+impl Command for FromTsv {
+    fn name(&self) -> &str {
+        "from tsv"
+    }
+
+    fn signature(&self) -> Signature {
+        Signature::build("from csv").switch(
+            "noheaders",
+            "don't treat the first row as column names",
+            Some('n'),
+        )
+    }
+
+    fn usage(&self) -> &str {
+        "Parse text as .csv and create table."
+    }
+
+    fn run(
+        &self,
+        _engine_state: &EngineState,
+        _stack: &mut Stack,
+        call: &Call,
+        input: PipelineData,
+    ) -> Result<nu_protocol::PipelineData, ShellError> {
+        from_tsv(call, input)
+    }
+}
+
+fn from_tsv(call: &Call, input: PipelineData) -> Result<PipelineData, ShellError> {
+    let name = call.head;
+
+    let noheaders = call.has_flag("noheaders");
+
+    from_delimited_data(noheaders, '\t', input, name)
+}
+
+#[cfg(test)]
+mod test {
+    use super::*;
+
+    #[test]
+    fn test_examples() {
+        use crate::test_examples;
+
+        test_examples(FromTsv {})
+    }
+}

crates/nu-engine/src/eval.rs

@@ -415,7 +415,7 @@ pub fn eval_subexpression(
                             // to be used later
                             // FIXME: the trimming of the end probably needs to live in a better place
 
-                            let mut s = input.collect_string();
+                            let mut s = input.collect_string("");
                             if s.ends_with('\n') {
                                 s.pop();
                             }

crates/nu-protocol/src/pipeline_data.rs

@@ -51,10 +51,10 @@ impl PipelineData {
         }
     }
 
-    pub fn collect_string(self) -> String {
+    pub fn collect_string(self, separator: &str) -> String {
         match self {
-            PipelineData::Value(v) => v.into_string("\n"),
-            PipelineData::Stream(s) => s.into_string("\n"),
+            PipelineData::Value(v) => v.into_string(separator),
+            PipelineData::Stream(s) => s.into_string(separator),
         }
     }
 

crates/nu-protocol/src/shell_error.rs

@@ -60,6 +60,10 @@ pub enum ShellError {
         right_span: Span,
     },
 
+    #[error("Delimiter error")]
+    #[diagnostic(code(nu::shell::delimiter_error), url(docsrs))]
+    DelimiterError(String, #[label("{0}")] Span),
+
     #[error("Incompatible parameters.")]
     #[diagnostic(code(nu::shell::incompatible_parameters), url(docsrs))]
     IncompatibleParametersSingle(String, #[label = "{0}"] Span),

crates/nu-protocol/src/value/stream.rs

@@ -20,12 +20,9 @@ pub struct ValueStream {
 
 impl ValueStream {
     pub fn into_string(self, separator: &str) -> String {
-        format!(
-            "[{}]",
-            self.map(|x: Value| x.into_string(", "))
-                .collect::<Vec<String>>()
-                .join(separator)
-        )
+        self.map(|x: Value| x.into_string(", "))
+            .collect::<Vec<String>>()
+            .join(separator)
     }
 
     pub fn from_stream(

src/main.rs

@@ -133,7 +133,7 @@ fn main() -> Result<()> {
             PipelineData::new(Span::unknown()),
         ) {
             Ok(pipeline_data) => {
-                println!("{}", pipeline_data.collect_string());
+                println!("{}", pipeline_data.collect_string("\n"));
             }
             Err(err) => {
                 let working_set = StateWorkingSet::new(&engine_state);
@@ -273,7 +273,7 @@ fn print_value(value: Value, engine_state: &EngineState) -> Result<(), ShellErro
                 &Call::new(),
                 value.into_pipeline_data(),
             )?;
-            table.collect_string()
+            table.collect_string("\n")
         }
         None => value.into_string(", "),
     };
@@ -323,7 +323,7 @@ fn update_prompt<'prompt>(
         &block,
         PipelineData::new(Span::unknown()),
     ) {
-        Ok(pipeline_data) => pipeline_data.collect_string(),
+        Ok(pipeline_data) => pipeline_data.collect_string(""),
         Err(err) => {
             let working_set = StateWorkingSet::new(engine_state);
             report_error(&working_set, &err);