Split document content into its own table to avoid having to load it each time. Move document creation into the model

2 files changed, 22 insertions, 15 deletions

diff --git a/archivist/cli.py b/archivist/cli.py
index b10e11b..31646aa 100644
--- a/archivist/cli.py
+++ b/archivist/cli.py
@@ -237,8 +237,10 @@ def doc():
 def add_doc(file, tags, create_tags, ignore_missing_tags):
     """Add a new document together with the given tags."""
 
-    from .model import Document, Tag, db, DocumentTag
+    from .model import db, Document
+
     import magic
+    mimetype = magic.from_file(file.name, mime=True)
 
     with db.atomic():
         if tags:
@@ -247,15 +249,7 @@ def add_doc(file, tags, create_tags, ignore_missing_tags):
             else:
                 tags = fetch_tags(tags, ignore_missing_tags)
 
-        mimetype = magic.from_file(file.name, mime=True)
-
-        doc = Document.create(content = file.read(),
-                file_type = mimetype,
-                original_path = file.name,
-                direction = Document.Direction.IN)
-
-        for t in tags:
-            DocumentTag.create(document = doc, tag = t)
+        Document.create_from_file(file, tags, direction = Document.Direction.IN, file_type = mimetype)
 
 @doc.command('find')
 @click.argument('tags', type=TAG, nargs=-1)
diff --git a/archivist/model.py b/archivist/model.py
index cbbc462..86c2cf1 100644
--- a/archivist/model.py
+++ b/archivist/model.py
@@ -63,25 +63,37 @@ class CompressedField(_CompressedField):
         return value if value is None else self.decompress(value)
 
 @table
+class DocumentContent(BaseModel):
+    blob = CompressedField()
+
+@table
 class Document(BaseModel):
     @unique
     class Direction(Enum):
         IN = 0
         OUT = 1
 
-    content = CompressedField()
     created = DateField(default=datetime.date.today)
     inserted = DateTimeField(default=datetime.datetime.now)
     description = CharField(null=True)
     original_path = CharField(null=True)
     file_type = CharField(null=True)
     direction = EnumField(Direction, null=True)
+    content = ForeignKeyField(DocumentContent, related_name = 'document')
 
     @classmethod
-    def matches(cls, prefix, value):
-        return query_pseudo_prefix(prefix, value) or Document.id << (
-                DocumentTag.select(DocumentTag.document)
-                .join(Tag, on=Tag.matches(prefix, value)))
+    @db.atomic()
+    def create_from_file(cls, from_file, tags, **kwargs):
+        content = DocumentContent.create(blob=from_file.read())
+
+        doc = cls.create(content = content,
+                original_path = from_file.name,
+                **kwargs)
+
+        for t in tags:
+            DocumentTag.create(document = doc, tag = t)
+
+        return doc
 
 @table
 class Prefix(BaseModel):
@@ -95,6 +107,7 @@ class Prefix(BaseModel):
     def create(cls, **query):
         inst = super().create(**query)
         Tag.create_prefix_default(inst)
+        return inst
 
     @property
     def default_tag(self):