From 0f3b0bcb6c196a9b93075a8abee897634c9c892f Mon Sep 17 00:00:00 2001
From: Robin Krahl <robin.krahl@ireas.org>
Date: Sat, 5 Sep 2020 13:19:01 +0200
Subject: Add tsv output format

This patch adds a new output format, tsv.  It uses the tsv crate to
generate the data and the serde_json crate to transform structs into
key-value pairs.

TODO: man page, changelog
---
 Cargo.lock    | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++
 Cargo.toml    |  3 +++
 src/args.rs   |  1 +
 src/output.rs | 59 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 113 insertions(+)

diff --git a/Cargo.lock b/Cargo.lock
index 2115b2e..5881f10 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -62,6 +62,24 @@ dependencies = [
  "constant_time_eq",
 ]
 
+[[package]]
+name = "bstr"
+version = "0.2.13"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "31accafdb70df7871592c058eca3985b71104e15ac32f64706022c58867da931"
+dependencies = [
+ "lazy_static",
+ "memchr",
+ "regex-automata",
+ "serde",
+]
+
+[[package]]
+name = "byteorder"
+version = "1.3.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "08c48aae112d48ed9f069b33538ea9e3e90aa263cfa3d1c24309612b1f7472de"
+
 [[package]]
 name = "cc"
 version = "1.0.50"
@@ -102,6 +120,28 @@ dependencies = [
  "lazy_static",
 ]
 
+[[package]]
+name = "csv"
+version = "1.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "00affe7f6ab566df61b4be3ce8cf16bc2576bca0963ceb0955e45d514bf9a279"
+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 = "directories"
 version = "3.0.1"
@@ -203,6 +243,7 @@ version = "0.3.3"
 dependencies = [
  "anyhow",
  "base32",
+ "csv",
  "directories",
  "envy",
  "libc",
@@ -346,6 +387,15 @@ dependencies = [
  "thread_local",
 ]
 
+[[package]]
+name = "regex-automata"
+version = "0.1.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ae1ded71d66a4a97f5e961fd0cb25a5f366a42a41570d16a763a69c092c26ae4"
+dependencies = [
+ "byteorder",
+]
+
 [[package]]
 name = "regex-syntax"
 version = "0.6.17"
diff --git a/Cargo.toml b/Cargo.toml
index a1c0e3b..2658b70 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -38,6 +38,9 @@ version = "1.0"
 [dependencies.base32]
 version = "0.4.0"
 
+[dependencies.csv]
+version = "1.1.0"
+
 [dependencies.envy]
 version = "0.4.1"
 
diff --git a/src/args.rs b/src/args.rs
index 3936738..f34de81 100644
--- a/src/args.rs
+++ b/src/args.rs
@@ -63,6 +63,7 @@ Enum! {
   OutputFormat, [
     Json => "json",
     Text => "text",
+    Tsv => "tsv",
   ]
 }
 
diff --git a/src/output.rs b/src/output.rs
index 37daf92..01dee2c 100644
--- a/src/output.rs
+++ b/src/output.rs
@@ -76,6 +76,7 @@ impl<T: fmt::Display + serde::Serialize> Output for Value<T> {
   fn format(&self, format: args::OutputFormat) -> anyhow::Result<String> {
     match format {
       args::OutputFormat::Json => get_json(&self.key, &self.value),
+      args::OutputFormat::Tsv => get_tsv_object(&self.value),
       args::OutputFormat::Text => Ok(self.value.to_string()),
     }
   }
@@ -103,6 +104,7 @@ impl<T: TableItem> Output for Table<T> {
   fn format(&self, format: args::OutputFormat) -> anyhow::Result<String> {
     match format {
       args::OutputFormat::Json => get_json(&self.key, &self.items),
+      args::OutputFormat::Tsv => get_tsv_list(&self.items),
       args::OutputFormat::Text => {
         if self.items.is_empty() {
           Ok(self.empty_message.clone())
@@ -192,3 +194,60 @@ fn get_json<T: serde::Serialize + ?Sized>(key: &str, value: &T) -> anyhow::Resul
   let _ = map.insert(key, value);
   serde_json::to_string_pretty(&map).context("Could not serialize output to JSON")
 }
+
+fn get_tsv_list<T: serde::Serialize>(items: &[T]) -> anyhow::Result<String> {
+  let mut writer = csv::WriterBuilder::new()
+    .delimiter(b'\t')
+    .from_writer(vec![]);
+  for item in items {
+    writer
+      .serialize(item)
+      .context("Could not serialize output to TSV")?;
+  }
+  String::from_utf8(writer.into_inner()?).context("Could not parse TSV output as UTF-8")
+}
+
+fn get_tsv_object<T: serde::Serialize>(value: T) -> anyhow::Result<String> {
+  let value = serde_json::to_value(&value).context("Could not serialize output")?;
+  get_tsv_list(&get_tsv_records(&[], value))
+}
+
+/// Converts an arbitrary value into a list of TSV records.
+///
+/// There are two cases:  Scalars are converted to a single value (without headers).  Arrays and
+/// objects are converted to a list of key-value pairs (with headers).  Nested arrays and objects
+/// are flattened and their keys are separated by dots.
+///
+/// `prefix` is the prefix to use for the keys of arrays and objects, or an empty slice if the
+/// given value is the top-level value.
+fn get_tsv_records(prefix: &[&str], value: serde_json::Value) -> Vec<serde_json::Value> {
+  use serde_json::Value;
+
+  let mut vec = Vec::new();
+  if (value.is_array() || value.is_object()) && prefix.is_empty() {
+    vec.push(Value::Array(vec!["key".into(), "value".into()]));
+  }
+
+  match value {
+    Value::Object(o) => {
+      for (key, value) in o {
+        vec.append(&mut get_tsv_records(&[prefix, &[&key]].concat(), value));
+      }
+    }
+    Value::Array(a) => {
+      for (idx, value) in a.into_iter().enumerate() {
+        let idx = idx.to_string();
+        vec.append(&mut get_tsv_records(&[prefix, &[&idx]].concat(), value));
+      }
+    }
+    _ => {
+      if prefix.is_empty() {
+        vec.push(value);
+      } else {
+        vec.push(Value::Array(vec![prefix.join(".").into(), value]));
+      }
+    }
+  }
+
+  vec
+}
-- 
cgit v1.2.1