From 54525d2e213ea62d89cb08ae29358a62056aa616 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Mike=20F=C3=A4hrmann?= <mike_faehrmann@web.de>
Date: Sat, 25 Jun 2022 16:45:34 +0200
Subject: [PATCH] [formatter] implement slice operator as format specifier

this allows using a slice operator alongside other (special) format
specifiers like J, to first join list elements to a string and then
trimming that with a slice.

{tags:J, /[:50]}
---
 docs/formatting.md      |  6 ++++++
 gallery_dl/formatter.py | 29 +++++++++++++++++++++++------
 test/test_formatter.py  | 20 +++++++++++++++++++-
 3 files changed, 48 insertions(+), 7 deletions(-)

diff --git a/docs/formatting.md b/docs/formatting.md
index f01e4193..b9c4355e 100644
--- a/docs/formatting.md
+++ b/docs/formatting.md
@@ -139,6 +139,12 @@ Format specifiers can be used for advanced formatting by using the options provi
     <td><code>{empty:?[/]/}</code></td>
     <td><code></code></td>
 </tr>
+<tr>
+    <td><code>[&lt;start&gt;:&lt;stop&gt;]</code></td>
+    <td>Applies a <a href="https://python-reference.readthedocs.io/en/latest/docs/brackets/slicing.html">Slicing</a> operation to the current value, similar to <a href="#field-names">Field Names</a></td>
+    <td><code>{foo:[1:-1]}</code></td>
+    <td><code>oo&nbsp;Ba</code></td>
+</tr>
 <tr>
     <td rowspan="2"><code>L&lt;maxlen&gt;/&lt;repl&gt;/</code></td>
     <td rowspan="2">Replaces the entire output with <code>&lt;repl&gt;</code> if its length exceeds <code>&lt;maxlen&gt;</code></td>
diff --git a/gallery_dl/formatter.py b/gallery_dl/formatter.py
index f2c2d331..107c8ed6 100644
--- a/gallery_dl/formatter.py
+++ b/gallery_dl/formatter.py
@@ -232,12 +232,7 @@ def parse_field_name(field_name):
             func = operator.itemgetter
             try:
                 if ":" in key:
-                    start, _, stop = key.partition(":")
-                    stop, _, step = stop.partition(":")
-                    start = int(start) if start else None
-                    stop = int(stop) if stop else None
-                    step = int(step) if step else None
-                    key = slice(start, stop, step)
+                    key = _slice(key)
             except TypeError:
                 pass  # key is an integer
 
@@ -246,6 +241,16 @@ def parse_field_name(field_name):
     return first, funcs
 
 
+def _slice(indices):
+    start, _, stop = indices.partition(":")
+    stop, _, step = stop.partition(":")
+    return slice(
+        int(start) if start else None,
+        int(stop) if stop else None,
+        int(step) if step else None,
+    )
+
+
 def parse_format_spec(format_spec, conversion):
     fmt = build_format_func(format_spec)
     if not conversion:
@@ -283,6 +288,8 @@ def build_format_func(format_spec):
         fmt = format_spec[0]
         if fmt == "?":
             return _parse_optional(format_spec)
+        if fmt == "[":
+            return _parse_slice(format_spec)
         if fmt == "L":
             return _parse_maxlen(format_spec)
         if fmt == "J":
@@ -305,6 +312,16 @@ def _parse_optional(format_spec):
     return optional
 
 
+def _parse_slice(format_spec):
+    indices, _, format_spec = format_spec.partition("]")
+    slice = _slice(indices[1:])
+    fmt = build_format_func(format_spec)
+
+    def apply_slice(obj):
+        return fmt(obj[slice])
+    return apply_slice
+
+
 def _parse_maxlen(format_spec):
     maxlen, replacement, format_spec = format_spec.split("/", 2)
     maxlen = text.parse_int(maxlen[1:])
diff --git a/test/test_formatter.py b/test/test_formatter.py
index dc818b00..5b8ca0ab 100644
--- a/test/test_formatter.py
+++ b/test/test_formatter.py
@@ -1,7 +1,7 @@
 #!/usr/bin/env python3
 # -*- coding: utf-8 -*-
 
-# Copyright 2021 Mike Fährmann
+# Copyright 2021-2022 Mike Fährmann
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License version 2 as
@@ -135,6 +135,21 @@ class TestFormatter(unittest.TestCase):
         self._run_test("{a[:50:2]}", v[:50:2])
         self._run_test("{a[::]}"   , v)
 
+        self._run_test("{a:[1:10]}"  , v[1:10])
+        self._run_test("{a:[-10:-1]}", v[-10:-1])
+        self._run_test("{a:[5:]}" , v[5:])
+        self._run_test("{a:[50:]}", v[50:])
+        self._run_test("{a:[:5]}" , v[:5])
+        self._run_test("{a:[:50]}", v[:50])
+        self._run_test("{a:[:]}"  , v)
+        self._run_test("{a:[1:10:2]}"  , v[1:10:2])
+        self._run_test("{a:[-10:-1:2]}", v[-10:-1:2])
+        self._run_test("{a:[5::2]}" , v[5::2])
+        self._run_test("{a:[50::2]}", v[50::2])
+        self._run_test("{a:[:5:2]}" , v[:5:2])
+        self._run_test("{a:[:50:2]}", v[:50:2])
+        self._run_test("{a:[::]}"   , v)
+
     def test_maxlen(self):
         v = self.kwdict["a"]
         self._run_test("{a:L5/foo/}" , "foo")
@@ -177,6 +192,9 @@ class TestFormatter(unittest.TestCase):
         # join-and-replace
         self._run_test("{l:J-/Rb/E/}", "a-E-c")
 
+        # join and slice
+        self._run_test("{l:J-/[1:-1]}", "-b-")
+
         # optional-and-maxlen
         self._run_test("{d[a]:?</>/L1/too long/}", "<too long>")
         self._run_test("{d[c]:?</>/L5/too long/}", "")