From 35854188daf6f2ec4ff77ff63e795b0c7d89da9b Mon Sep 17 00:00:00 2001
From: Drew DeVault <sir@cmpwn.com>
Date: Fri, 21 Sep 2018 18:21:32 -0400
Subject: [PATCH] Implement tree annotation

---
 gitsrht/app.py              |  2 +
 gitsrht/blueprints/repo.py  |  6 +--
 gitsrht/git.py              | 91 ++++++++++++++++++++++++++++++++++++-
 gitsrht/templates/tree.html | 11 +++--
 4 files changed, 103 insertions(+), 7 deletions(-)

diff --git a/gitsrht/app.py b/gitsrht/app.py
index 34c5207..2376a51 100644
--- a/gitsrht/app.py
+++ b/gitsrht/app.py
@@ -1,3 +1,4 @@
+import humanize
 import stat
 from flask import session
 from srht.flask import SrhtFlask
@@ -43,6 +44,7 @@ class GitApp(SrhtFlask):
             return {
                 "commit_time": commit_time,
                 "trim_commit": trim_commit,
+                "humanize": humanize,
                 "stat": stat,
                 "notice": notice
             }
diff --git a/gitsrht/blueprints/repo.py b/gitsrht/blueprints/repo.py
index 054afc6..e9e5651 100644
--- a/gitsrht/blueprints/repo.py
+++ b/gitsrht/blueprints/repo.py
@@ -4,7 +4,7 @@ from flask import Blueprint, render_template, abort
 from flask_login import current_user
 from gitsrht.access import get_repo, has_access, UserAccess
 from gitsrht.redis import redis
-from gitsrht.git import CachedRepository, commit_time
+from gitsrht.git import CachedRepository, commit_time, annotate_tree
 from gitsrht.types import User, Repository
 from srht.config import cfg
 from srht.markdown import markdown
@@ -83,8 +83,8 @@ def tree(owner, repo, branch, path):
     if not branch:
         abort(404)
     commit = git_repo.get(branch.target)
-    # TODO: annotate tree with latest commit for each file (and cache it)
-    tree = [entry for entry in commit.tree]
+
+    tree = annotate_tree(git_repo, commit)
     tree = sorted(tree, key=lambda e: e.name)
     # TODO: follow path
     return render_template("tree.html", view="tree",
diff --git a/gitsrht/git.py b/gitsrht/git.py
index 8c0a5a4..acfb6ab 100644
--- a/gitsrht/git.py
+++ b/gitsrht/git.py
@@ -1,6 +1,9 @@
+from collections import deque
 from datetime import datetime, timedelta, timezone
-from pygit2 import Repository, Tag
 from functools import lru_cache
+from gitsrht.redis import redis
+from pygit2 import Repository, Tag
+import json
 
 def trim_commit(msg):
     if "\n" not in msg:
@@ -39,3 +42,89 @@ class _CachedRepository(Repository):
             branch = list(self.branches.local)[0]
             branch = self.branches.get(branch)
         return branch
+
+class AnnotatedTreeEntry:
+    def __init__(self, repo, entry):
+        self._entry = entry
+        self._repo = repo
+        if entry:
+            self.id = entry.id.hex
+            self.name = entry.name
+            self.type = entry.type
+            self.filemode = entry.filemode
+
+    def fetch_blob(self):
+        if self.type == "tree":
+            self.tree = self._repo.get(self.id)
+        else:
+            self.blob = self._repo.get(self.id)
+        return self
+
+    def serialize(self):
+        return {
+            "id": self.id,
+            "name": self.name,
+            "type": self.type,
+            "filemode": self.filemode,
+            "commit": (self.commit.id.hex
+                if hasattr(self, "commit") and self.commit else None),
+        }
+
+    @staticmethod
+    def deserialize(res, repo):
+        _id = res["id"]
+        self = AnnotatedTreeEntry(repo, None)
+        self.id = res["id"]
+        self.name = res["name"]
+        self.type = res["type"]
+        self.filemode = res["filemode"]
+        self.commit = repo.get(res["commit"]) if "commit" in res else None
+        return self
+
+    def __hash__(self):
+        return hash(f"{self.id}:{self.name}")
+
+    def __eq__(self, other):
+        return self.id == other.id and self.name == other.name
+
+    def __repr__(self):
+        return f"<AnnotatedTreeEntry {self.name} {self.id}>"
+
+def annotate_tree(repo, commit):
+    key = f"git.sr.ht:git:tree:{commit.tree.id.hex}"
+    cache = redis.get(key)
+    if cache:
+        cache = json.loads(cache.decode())
+        return [AnnotatedTreeEntry.deserialize(
+            e, repo).fetch_blob() for e in cache.values()]
+
+    tree = { entry.id.hex: AnnotatedTreeEntry(
+        repo, entry) for entry in commit.tree }
+
+    parents = deque(commit.parents)
+    left_tree = set(v for v in tree.values())
+    unfinished = set(left_tree)
+    if not any(commit.parents):
+        return [entry.fetch_blob() for entry in tree.values()]
+    parent = commit.parents[0]
+
+    while any(unfinished):
+        right_tree = { entry.id.hex: AnnotatedTreeEntry(repo, entry)
+                for entry in parent.tree }
+        right_tree = set(v for v in right_tree.values())
+        diff = left_tree - right_tree
+        for entry in diff:
+            if entry.id in tree:
+                tree[entry.id].commit = commit
+        unfinished = unfinished - diff
+        left_tree = right_tree
+        commit = parent
+        if not any(commit.parents):
+            break
+        parent = commit.parents[0]
+
+    cache = {entry.name: entry.serialize() for entry in tree.values()}
+    cache = json.dumps(cache)
+    redis.setex(key, cache, timedelta(days=30))
+
+    return [entry.fetch_blob() for entry in tree.values()]
diff --git a/gitsrht/templates/tree.html b/gitsrht/templates/tree.html
index 3e79a26..d2d4df0 100644
--- a/gitsrht/templates/tree.html
+++ b/gitsrht/templates/tree.html
@@ -27,11 +27,16 @@
         <a href="#">{{entry.name}}{% if entry.type == "tree" %}/{% endif %}</a>
       </div>
       <div class="commit">
-        <a href="#">{{trim_commit(commit.message)}}</a> &mdash;
-        {{ commit_time(commit) | date }}
+        {% if entry.commit %}
+        <a href="#">{{trim_commit(entry.commit.message)}}</a> &mdash;
+        {{ commit_time(entry.commit) | date }}
+        {% endif %}
       </div>
       <div class="size">
-        1234 bytes
+        {% if entry.type == "blob" %}
+        {{humanize.naturalsize(entry.blob.size,
+          binary=True).replace("Byte", "byte")}}
+        {% endif %}
       </div>
       {% endfor %}
       </div>
-- 
2.38.4