Add an output pager.

Fixes #50.
Requires pallets/click#1572.

sno/diff_output.py

@@ -11,7 +11,10 @@
 import click
 
 from . import gpkg
-from .output_util import dump_json_output, resolve_output_path
+from .output_util import (
+    dump_json_output,
+    resolve_output_path,
+)
 
 
 @contextlib.contextmanager
@@ -50,82 +53,82 @@ def diff_output_text(*, output_path, **kwargs):
     In particular, geometry WKT is abbreviated and null values are represented
     by a unicode "␀" character.
     """
-    fp = resolve_output_path(output_path)
-    pecho = {'file': fp, 'color': fp.isatty()}
     if isinstance(output_path, Path) and output_path.is_dir():
         raise click.BadParameter(
             "Directory is not valid for --output with --text", param_hint="--output"
         )
 
     def _out(dataset, diff):
-        path = dataset.path
-        pk_field = dataset.primary_key
-        prefix = f"{path}:"
-        repr_excl = [pk_field]
-
-        for k, (v_old, v_new) in diff["META"].items():
-            click.secho(
-                f"--- {prefix}meta/{k}\n+++ {prefix}meta/{k}", bold=True, **pecho
-            )
-
-            s_old = set(v_old.items())
-            s_new = set(v_new.items())
-
-            diff_add = dict(s_new - s_old)
-            diff_del = dict(s_old - s_new)
-            all_keys = set(diff_del.keys()) | set(diff_add.keys())
-
-            for k in all_keys:
-                if k in diff_del:
-                    click.secho(
-                        text_row({k: diff_del[k]}, prefix="- ", exclude=repr_excl),
-                        fg="red",
-                        **pecho,
-                    )
-                if k in diff_add:
-                    click.secho(
-                        text_row({k: diff_add[k]}, prefix="+ ", exclude=repr_excl),
-                        fg="green",
-                        **pecho,
-                    )
-
-        prefix = f"{path}:{pk_field}="
+        with resolve_output_path(output_path) as fp:
+            pecho = {'file': fp}
+            path = dataset.path
+            pk_field = dataset.primary_key
+            prefix = f"{path}:"
+            repr_excl = [pk_field]
+
+            for k, (v_old, v_new) in diff["META"].items():
+                click.secho(
+                    f"--- {prefix}meta/{k}\n+++ {prefix}meta/{k}", bold=True, **pecho
+                )
 
-        for k, v_old in diff["D"].items():
-            click.secho(f"--- {prefix}{k}", bold=True, **pecho)
-            click.secho(
-                text_row(v_old, prefix="- ", exclude=repr_excl), fg="red", **pecho
-            )
+                s_old = set(v_old.items())
+                s_new = set(v_new.items())
+
+                diff_add = dict(s_new - s_old)
+                diff_del = dict(s_old - s_new)
+                all_keys = set(diff_del.keys()) | set(diff_add.keys())
+
+                for k in all_keys:
+                    if k in diff_del:
+                        click.secho(
+                            text_row({k: diff_del[k]}, prefix="- ", exclude=repr_excl),
+                            fg="red",
+                            **pecho,
+                        )
+                    if k in diff_add:
+                        click.secho(
+                            text_row({k: diff_add[k]}, prefix="+ ", exclude=repr_excl),
+                            fg="green",
+                            **pecho,
+                        )
+
+            prefix = f"{path}:{pk_field}="
+
+            for k, v_old in diff["D"].items():
+                click.secho(f"--- {prefix}{k}", bold=True, **pecho)
+                click.secho(
+                    text_row(v_old, prefix="- ", exclude=repr_excl), fg="red", **pecho
+                )
 
-        for o in diff["I"]:
-            click.secho(f"+++ {prefix}{o[pk_field]}", bold=True, **pecho)
-            click.secho(
-                text_row(o, prefix="+ ", exclude=repr_excl), fg="green", **pecho
-            )
+            for o in diff["I"]:
+                click.secho(f"+++ {prefix}{o[pk_field]}", bold=True, **pecho)
+                click.secho(
+                    text_row(o, prefix="+ ", exclude=repr_excl), fg="green", **pecho
+                )
 
-        for _, (v_old, v_new) in diff["U"].items():
-            click.secho(
-                f"--- {prefix}{v_old[pk_field]}\n+++ {prefix}{v_new[pk_field]}",
-                bold=True,
-                **pecho,
-            )
+            for _, (v_old, v_new) in diff["U"].items():
+                click.secho(
+                    f"--- {prefix}{v_old[pk_field]}\n+++ {prefix}{v_new[pk_field]}",
+                    bold=True,
+                    **pecho,
+                )
 
-            s_old = set(v_old.items())
-            s_new = set(v_new.items())
+                s_old = set(v_old.items())
+                s_new = set(v_new.items())
 
-            diff_add = dict(s_new - s_old)
-            diff_del = dict(s_old - s_new)
-            all_keys = sorted(set(diff_del.keys()) | set(diff_add.keys()))
+                diff_add = dict(s_new - s_old)
+                diff_del = dict(s_old - s_new)
+                all_keys = sorted(set(diff_del.keys()) | set(diff_add.keys()))
 
-            for k in all_keys:
-                if k in diff_del:
-                    rk = text_row({k: diff_del[k]}, prefix="- ", exclude=repr_excl)
-                    if rk:
-                        click.secho(rk, fg="red", **pecho)
-                if k in diff_add:
-                    rk = text_row({k: diff_add[k]}, prefix="+ ", exclude=repr_excl)
-                    if rk:
-                        click.secho(rk, fg="green", **pecho)
+                for k in all_keys:
+                    if k in diff_del:
+                        rk = text_row({k: diff_del[k]}, prefix="- ", exclude=repr_excl)
+                        if rk:
+                            click.secho(rk, fg="red", **pecho)
+                    if k in diff_add:
+                        rk = text_row({k: diff_add[k]}, prefix="+ ", exclude=repr_excl)
+                        if rk:
+                            click.secho(rk, fg="green", **pecho)
 
     yield _out
 
@@ -367,20 +370,19 @@ def diff_output_html(*, output_path, repo, base, target, dataset_count, **kwargs
 
         if not output_path:
             output_path = Path(repo.path) / "DIFF.html"
-        fo = resolve_output_path(output_path)
-
-        # Read all the geojson back in, and stick them in a dict
-        all_datasets_geojson = {}
-        for filename in os.listdir(tempdir):
-            with open(tempdir / filename) as json_file:
-                all_datasets_geojson[os.path.splitext(filename)[0]] = json.load(
-                    json_file
+        with resolve_output_path(output_path) as fo:
+            # Read all the geojson back in, and stick them in a dict
+            all_datasets_geojson = {}
+            for filename in os.listdir(tempdir):
+                with open(tempdir / filename) as json_file:
+                    all_datasets_geojson[os.path.splitext(filename)[0]] = json.load(
+                        json_file
+                    )
+            fo.write(
+                template.substitute(
+                    {"title": title, "geojson_data": json.dumps(all_datasets_geojson)}
                 )
-        fo.write(
-            template.substitute(
-                {"title": title, "geojson_data": json.dumps(all_datasets_geojson)}
             )
-        )
-    if fo != sys.stdout:
-        fo.close()
-        webbrowser.open_new(f"file://{output_path.resolve()}")
+        if fo != sys.stdout:
+            fo.close()
+            webbrowser.open_new(f"file://{output_path.resolve()}")

sno/output_util.py

@@ -1,6 +1,14 @@
 import io
 import json
+import os
+import shutil
 import sys
+import threading
+from contextlib import closing, contextmanager
+from queue import Queue, Empty
+
+import click
+from click._compat import should_strip_ansi
 
 JSON_PARAMS = {
     "compact": {},
@@ -14,35 +22,54 @@ def dump_json_output(output, output_path, json_style="pretty"):
     Dumps the output to JSON in the output file.
     """
 
-    fp = resolve_output_path(output_path)
-
-    if json_style == 'pretty' and fp == sys.stdout and fp.isatty():
-        # Add syntax highlighting
-        from pygments import highlight
-        from pygments.lexers import JsonLexer
-        from pygments.formatters import TerminalFormatter
+    with resolve_output_path(output_path) as fp:
+        if json_style == 'pretty' and not should_strip_ansi(fp):
+            # Add syntax highlighting
+            from pygments import highlight
+            from pygments.lexers import JsonLexer
+            from pygments.formatters import TerminalFormatter
 
-        dumped = json.dumps(output, **JSON_PARAMS[json_style])
-        highlighted = highlight(dumped.encode(), JsonLexer(), TerminalFormatter())
-        fp.write(highlighted)
-    else:
-        json.dump(output, fp, **JSON_PARAMS[json_style])
+            dumped = json.dumps(output, **JSON_PARAMS[json_style])
+            highlighted = highlight(dumped.encode(), JsonLexer(), TerminalFormatter())
+            fp.write(highlighted)
+        else:
+            json.dump(output, fp, **JSON_PARAMS[json_style])
 
 
-def resolve_output_path(output_path):
+@contextmanager
+def resolve_output_path(output_path, allow_pager=True):
     """
-    Takes a path-ish thing, and returns the appropriate writable file-like object.
+    Context manager.
+
+    Takes a path-ish thing, and yields the appropriate writable file-like object.
     The path-ish thing could be:
       * a pathlib.Path object
       * a file-like object
       * the string '-' or None (both will return sys.stdout)
+
+    If the file is not stdout, it will be closed when exiting the context manager.
+
+    If allow_pager=True (the default) and the file is stdout, this will attempt to use a
+    pager to display long output.
     """
+
     if isinstance(output_path, io.IOBase):
-        return output_path
+        # Make this contextmanager re-entrant
+        yield output_path
     elif (not output_path) or output_path == "-":
-        return sys.stdout
+        if allow_pager and get_input_mode() == InputMode.INTERACTIVE:
+            pager_cmd = (
+                os.environ.get('SNO_PAGER') or os.environ.get('PAGER') or DEFAULT_PAGER
+            )
+
+            with _push_environment('PAGER', pager_cmd):
+                with click.get_pager_file() as pager:
+                    yield pager
+        else:
+            yield sys.stdout
     else:
-        return output_path.open("w")
+        with closing(output_path.open("w")) as f:
+            yield f
 
 
 class InputMode:
@@ -69,3 +96,25 @@ def is_empty_stream(stream):
             return True
         stream.seek(pos)
     return False
+
+
+def _setenv(k, v):
+    if v is None:
+        os.environ.pop(k, None)
+    else:
+        os.environ[k] = v
+
+
+@contextmanager
+def _push_environment(k, v):
+    orig = os.environ.get(k)
+    _setenv(k, v)
+    try:
+        yield
+    finally:
+        _setenv(k, orig)
+
+
+DEFAULT_PAGER = shutil.which('less')
+if DEFAULT_PAGER:
+    DEFAULT_PAGER += ' --quit-if-one-screen --no-init -R'

sno/show.py

@@ -92,24 +92,24 @@ def patch_output_text(*, target, output_path, **kwargs):
     by a unicode "␀" character.
     """
     commit = target.head_commit
-    fp = resolve_output_path(output_path)
-    pecho = {'file': fp, 'color': fp.isatty()}
-    with diff.diff_output_text(output_path=fp, **kwargs) as diff_writer:
-        author = commit.author
-        author_time_utc = datetime.fromtimestamp(author.time, timezone.utc)
-        author_timezone = timezone(timedelta(minutes=author.offset))
-        author_time_in_author_timezone = author_time_utc.astimezone(author_timezone)
-
-        click.secho(f'commit {commit.hex}', fg='yellow')
-        click.secho(f'Author: {author.name} <{author.email}>', **pecho)
-        click.secho(
-            f'Date:   {author_time_in_author_timezone.strftime("%c %z")}', **pecho
-        )
-        click.secho(**pecho)
-        for line in commit.message.splitlines():
-            click.secho(f'    {line}', **pecho)
-        click.secho(**pecho)
-        yield diff_writer
+    with resolve_output_path(output_path) as fp:
+        pecho = {'file': fp}
+        with diff.diff_output_text(output_path=fp, **kwargs) as diff_writer:
+            author = commit.author
+            author_time_utc = datetime.fromtimestamp(author.time, timezone.utc)
+            author_timezone = timezone(timedelta(minutes=author.offset))
+            author_time_in_author_timezone = author_time_utc.astimezone(author_timezone)
+
+            click.secho(f'commit {commit.hex}', fg='yellow', **pecho)
+            click.secho(f'Author: {author.name} <{author.email}>', **pecho)
+            click.secho(
+                f'Date:   {author_time_in_author_timezone.strftime("%c %z")}', **pecho
+            )
+            click.secho(**pecho)
+            for line in commit.message.splitlines():
+                click.secho(f'    {line}', **pecho)
+            click.secho(**pecho)
+            yield diff_writer
 
 
 @contextlib.contextmanager