Initial importv0.1.0

1 files changed, 122 insertions, 0 deletions

diff --git a/src/Cepis/Csv.cs b/src/Cepis/Csv.cs
new file mode 100644
index 0000000..ca11558
--- /dev/null
+++ b/src/Cepis/Csv.cs
@@ -0,0 +1,122 @@
+using System.Text;
+
+namespace Cepis;
+
+/// <summary>
+/// An RFC 4180 reader for the batch endpoint.
+/// </summary>
+/// <remarks>
+/// Values may contain commas, quotes and newlines, so they arrive
+/// double-quoted with <c>""</c> escaping; splitting on commas would corrupt
+/// exactly the addresses that need care. A leading byte order mark is dropped,
+/// and CR, LF and CRLF all end a record. The BCL ships no CSV reader, and the
+/// shape this endpoint returns is small enough not to justify a dependency.
+/// </remarks>
+internal static class Csv
+{
+    /// <summary>
+    /// Parse a CSV document whose first record is the header.
+    /// </summary>
+    /// <remarks>
+    /// Each record becomes a dictionary from column name to value. A record
+    /// with fewer fields than the header has the missing ones filled with
+    /// <c>""</c>; extra fields are dropped.
+    /// </remarks>
+    /// <param name="text">the document</param>
+    /// <returns>one dictionary per data record, in order</returns>
+    internal static List<Dictionary<string, string>> Parse(string text)
+    {
+        var records = Split(text);
+        var rows = new List<Dictionary<string, string>>();
+        if (records.Count == 0)
+        {
+            return rows;
+        }
+
+        var header = records[0];
+        for (var r = 1; r < records.Count; r++)
+        {
+            var record = records[r];
+            var row = new Dictionary<string, string>(header.Count, StringComparer.Ordinal);
+            for (var i = 0; i < header.Count; i++)
+            {
+                row[header[i]] = i < record.Count ? record[i] : "";
+            }
+
+            rows.Add(row);
+        }
+
+        return rows;
+    }
+
+    private static List<List<string>> Split(string text)
+    {
+        var records = new List<List<string>>();
+        var record = new List<string>();
+        var field = new StringBuilder();
+        var quoted = false;
+        var pending = false;
+
+        var i = text.Length > 0 && text[0] == '\uFEFF' ? 1 : 0;
+        for (; i < text.Length; i++)
+        {
+            var c = text[i];
+            if (quoted)
+            {
+                if (c != '"')
+                {
+                    field.Append(c);
+                }
+                else if (i + 1 < text.Length && text[i + 1] == '"')
+                {
+                    field.Append('"');
+                    i++;
+                }
+                else
+                {
+                    quoted = false;
+                }
+
+                continue;
+            }
+
+            switch (c)
+            {
+                case '"':
+                    quoted = true;
+                    pending = true;
+                    break;
+                case ',':
+                    record.Add(field.ToString());
+                    field.Clear();
+                    pending = true;
+                    break;
+                case '\r':
+                case '\n':
+                    if (c == '\r' && i + 1 < text.Length && text[i + 1] == '\n')
+                    {
+                        i++;
+                    }
+
+                    record.Add(field.ToString());
+                    field.Clear();
+                    records.Add(record);
+                    record = [];
+                    pending = false;
+                    break;
+                default:
+                    field.Append(c);
+                    pending = true;
+                    break;
+            }
+        }
+
+        if (pending || field.Length > 0)
+        {
+            record.Add(field.ToString());
+            records.Add(record);
+        }
+
+        return records;
+    }
+}