Started work on cli, which also meant adding setup.py

I'm using click, and click recommends using a setup.py - so I've added one of
those. I also refactored code into a new datasite package. It's not quite
deploying to now properly at the moment though - I seem to have messed up the
path handling a bit.

Also snuck in a new template for the "Row" view.

Refs #40
This commit is contained in:
Simon Willison 2017-10-27 00:08:24 -07:00
commit 1592fd0419
15 changed files with 602 additions and 523 deletions

0
datasite/__init__.py Normal file
View file

420
datasite/app.py Normal file
View file

@ -0,0 +1,420 @@
from sanic import Sanic
from sanic import response
from sanic.exceptions import NotFound
from sanic.views import HTTPMethodView
from sanic_jinja2 import SanicJinja2
from jinja2 import FileSystemLoader
import click
import sqlite3
from contextlib import contextmanager
from pathlib import Path
from functools import wraps
import urllib.parse
import json
import base64
import hashlib
import sys
import time
app_root = Path(__file__).parent.parent
BUILD_METADATA = 'build-metadata.json'
DB_GLOBS = ('*.db', '*.sqlite', '*.sqlite3')
HASH_BLOCK_SIZE = 1024 * 1024
SQL_TIME_LIMIT_MS = 1000
conns = {}
app = Sanic(__name__)
jinja = SanicJinja2(
app,
loader=FileSystemLoader([
str(app_root / 'datasite' / 'templates')
])
)
def get_conn(name):
if name not in conns:
info = ensure_build_metadata()[name]
conns[name] = sqlite3.connect(
'file:{}?immutable=1'.format(info['file']),
uri=True
)
conns[name].row_factory = sqlite3.Row
conns[name].text_factory = lambda x: str(x, 'utf-8', 'replace')
return conns[name]
def ensure_build_metadata(regenerate=False):
build_metadata = app_root / BUILD_METADATA
if build_metadata.exists() and not regenerate:
return json.loads(build_metadata.read_text())
print('Building metadata... path={}'.format(build_metadata))
metadata = {}
for glob in DB_GLOBS:
for path in app_root.glob(glob):
print(' globbing, path={}'.format(path))
name = path.stem
if name in metadata:
raise Exception('Multiple files with same stem %s' % name)
# Calculate hash, efficiently
m = hashlib.sha256()
with path.open('rb') as fp:
while True:
data = fp.read(HASH_BLOCK_SIZE)
if not data:
break
m.update(data)
# List tables and their row counts
tables = {}
with sqlite3.connect('file:{}?immutable=1'.format(path.name), uri=True) as conn:
conn.row_factory = sqlite3.Row
table_names = [
r['name']
for r in conn.execute('select * from sqlite_master where type="table"')
]
for table in table_names:
tables[table] = conn.execute('select count(*) from "{}"'.format(table)).fetchone()[0]
metadata[name] = {
'hash': m.hexdigest(),
'file': path.name,
'tables': tables,
}
build_metadata.write_text(json.dumps(metadata, indent=4))
return metadata
class BaseView(HTTPMethodView):
template = None
def redirect(self, request, path):
if request.query_string:
path = '{}?{}'.format(
path, request.query_string
)
r = response.redirect(path)
r.headers['Link'] = '<{}>; rel=preload'.format(path)
return r
async def get(self, request, db_name, **kwargs):
name, hash, should_redirect = resolve_db_name(db_name, **kwargs)
if should_redirect:
return self.redirect(request, should_redirect)
return await self.view_get(request, name, hash, **kwargs)
async def view_get(self, request, name, hash, **kwargs):
try:
as_json = kwargs.pop('as_json')
except KeyError:
as_json = False
extra_template_data = {}
start = time.time()
try:
data, extra_template_data = self.data(
request, name, hash, **kwargs
)
except sqlite3.OperationalError as e:
data = {
'ok': False,
'error': str(e),
}
end = time.time()
data['took_ms'] = (end - start) * 1000
if as_json:
# Special case for .jsono extension
if as_json == '.jsono':
columns = data.get('columns')
rows = data.get('rows')
if rows and columns:
data['rows'] = [
dict(zip(columns, row))
for row in rows
]
r = response.HTTPResponse(
json.dumps(
data, cls=CustomJSONEncoder
),
content_type='application/json',
headers={
'Access-Control-Allow-Origin': '*'
}
)
else:
context = {**data, **dict(
extra_template_data()
if callable(extra_template_data)
else extra_template_data
)}
r = jinja.render(
self.template,
request,
**context,
)
# Set far-future cache expiry
r.headers['Cache-Control'] = 'max-age={}'.format(
365 * 24 * 60 * 60
)
return r
@app.route('/')
async def index(request, sql=None):
databases = []
for key, info in ensure_build_metadata().items():
database = {
'name': key,
'hash': info['hash'],
'path': '{}-{}'.format(key, info['hash'][:7]),
'tables_truncated': sorted(
info['tables'].items(),
key=lambda p: p[1],
reverse=True
)[:5],
'tables_count': len(info['tables'].items()),
'tables_more': len(info['tables'].items()) > 5,
'total_rows': sum(info['tables'].values()),
}
databases.append(database)
return jinja.render(
'index.html',
request,
databases=databases,
)
@app.route('/favicon.ico')
async def favicon(request):
return response.text('')
class DatabaseView(BaseView):
template = 'database.html'
def data(self, request, name, hash):
conn = get_conn(name)
sql = request.args.get('sql') or 'select * from sqlite_master'
with sqlite_timelimit(conn, SQL_TIME_LIMIT_MS):
rows = conn.execute(sql)
columns = [r[0] for r in rows.description]
return {
'database': name,
'rows': rows,
'columns': columns,
}, {
'database_hash': hash,
}
class DatabaseDownload(BaseView):
async def view_get(self, request, name, hash, **kwargs):
filepath = ensure_build_metadata()[name]['file']
return await response.file_stream(
filepath, headers={
'Content-Disposition': 'attachment; filename="{}"'.format(filepath)
}
)
class TableView(BaseView):
template = 'table.html'
def data(self, request, name, hash, table):
conn = get_conn(name)
table = urllib.parse.unquote_plus(table)
if request.args:
where_clause, params = build_where_clause(request.args)
sql = 'select * from "{}" where {} limit 50'.format(
table, where_clause
)
else:
sql = 'select * from "{}" limit 50'.format(table)
params = []
with sqlite_timelimit(conn, SQL_TIME_LIMIT_MS):
rows = conn.execute(sql, params)
columns = [r[0] for r in rows.description]
rows = list(rows)
pks = pks_for_table(conn, table)
info = ensure_build_metadata()
total_rows = info[name]['tables'].get(table)
return {
'database': name,
'table': table,
'rows': rows,
'total_rows': total_rows,
'columns': columns,
'primary_keys': pks,
}, lambda: {
'database_hash': hash,
'row_link': lambda row: path_from_row_pks(row, pks),
}
class RowView(BaseView):
template = 'row.html'
def data(self, request, name, hash, table, pk_path):
conn = get_conn(name)
pk_values = compound_pks_from_path(pk_path)
pks = pks_for_table(conn, table)
wheres = [
'"{}"=?'.format(pk)
for pk in pks
]
sql = 'select * from "{}" where {}'.format(
table, ' AND '.join(wheres)
)
rows = conn.execute(sql, pk_values)
columns = [r[0] for r in rows.description]
pks = pks_for_table(conn, table)
rows = list(rows)
if not rows:
raise NotFound('Record not found: {}'.format(pk_values))
return {
'database': name,
'table': table,
'rows': rows,
'columns': columns,
'primary_keys': pks,
}, {
'database_hash': hash,
'row_link': None,
}
app.add_route(DatabaseView.as_view(), '/<db_name:[^/\.]+?><as_json:(.jsono?)?$>')
app.add_route(DatabaseDownload.as_view(), '/<db_name:[^/]+?><as_db:(\.db)$>')
app.add_route(TableView.as_view(), '/<db_name:[^/]+>/<table:[^/]+?><as_json:(.jsono?)?$>')
app.add_route(RowView.as_view(), '/<db_name:[^/]+>/<table:[^/]+?>/<pk_path:[^/]+?><as_json:(.jsono?)?$>')
def resolve_db_name(db_name, **kwargs):
databases = ensure_build_metadata()
hash = None
name = None
if '-' in db_name:
# Might be name-and-hash, or might just be
# a name with a hyphen in it
name, hash = db_name.rsplit('-', 1)
if name not in databases:
# Try the whole name
name = db_name
hash = None
else:
name = db_name
# Verify the hash
try:
info = databases[name]
except KeyError:
raise NotFound('Database not found: {}'.format(name))
expected = info['hash'][:7]
if expected != hash:
should_redirect = '/{}-{}'.format(
name, expected,
)
if 'table' in kwargs:
should_redirect += '/' + kwargs['table']
if 'as_json' in kwargs:
should_redirect += kwargs['as_json']
if 'as_db' in kwargs:
should_redirect += kwargs['as_db']
return name, expected, should_redirect
return name, expected, None
def compound_pks_from_path(path):
return [
urllib.parse.unquote_plus(b) for b in path.split(',')
]
def pks_for_table(conn, table):
rows = [
row for row in conn.execute(
'PRAGMA table_info("{}")'.format(table)
).fetchall()
if row[-1]
]
rows.sort(key=lambda row: row[-1])
return [str(r[1]) for r in rows]
def path_from_row_pks(row, pks):
if not pks:
return ''
bits = []
for pk in pks:
bits.append(
urllib.parse.quote_plus(str(row[pk]))
)
return ','.join(bits)
def build_where_clause(args):
sql_bits = []
for key, values in args.items():
if '__' in key:
column, lookup = key.rsplit('__', 1)
else:
column = key
lookup = 'exact'
template = {
'exact': '"{}" = ?',
'contains': '"{}" like ?',
'endswith': '"{}" like ?',
'startswith': '"{}" like ?',
'gt': '"{}" > ?',
'gte': '"{}" >= ?',
'lt': '"{}" < ?',
'lte': '"{}" <= ?',
'glob': '"{}" glob ?',
'like': '"{}" like ?',
}[lookup]
value = values[0]
value_convert = {
'contains': lambda s: '%{}%'.format(s),
'endswith': lambda s: '%{}'.format(s),
'startswith': lambda s: '{}%'.format(s),
}.get(lookup, lambda s: s)
converted = value_convert(value)
sql_bits.append(
(template.format(column), converted)
)
sql_bits.sort(key=lambda p: p[0])
where_clause = ' and '.join(p[0] for p in sql_bits)
params = [p[1] for p in sql_bits]
return where_clause, params
class CustomJSONEncoder(json.JSONEncoder):
def default(self, obj):
if isinstance(obj, sqlite3.Row):
return tuple(obj)
if isinstance(obj, sqlite3.Cursor):
return list(obj)
if isinstance(obj, bytes):
# Does it encode to utf8?
try:
return obj.decode('utf8')
except UnicodeDecodeError:
return {
'$base64': True,
'encoded': base64.b64encode(obj).decode('latin1'),
}
return json.JSONEncoder.default(self, obj)
@contextmanager
def sqlite_timelimit(conn, ms):
deadline = time.time() + (ms / 1000)
def handler():
if time.time() >= deadline:
return 1
conn.set_progress_handler(handler, 10000)
yield
conn.set_progress_handler(None, 10000)

24
datasite/cli.py Normal file
View file

@ -0,0 +1,24 @@
import click
from .app import app, ensure_build_metadata
@click.group()
def cli():
"""
Datasite!
"""
@cli.command()
def build():
ensure_build_metadata(True)
@cli.command()
@click.argument('files', type=click.Path(exists=True), nargs=-1)
@click.option('-h', '--host', default='0.0.0.0')
@click.option('-p', '--port', default=8001)
@click.option('--debug', is_flag=True)
def serve(files, host, port, debug):
'''Serve up specified database files with a web UI'''
click.echo('Serve! files={} on port {}'.format(files, port))
app.run(host=host, port=port, debug=debug)

View file

@ -0,0 +1,19 @@
<!DOCTYPE html>
<html>
<head>
<title>{% block title %}{% endblock %}</title>
<style>
th {
text-align: left;
}
</style>
{% block extra_head %}{% endblock %}
</head>
<body>
{% if error %}
<div style="padding: 1em; margin: 1em; border: 3px solid red;">{{ error }}</div>
{% endif %}
{% block content %}
{% endblock %}
</body>
</html>

View file

@ -0,0 +1,35 @@
{% extends "base.html" %}
{% block title %}{{ database }}{% endblock %}
{% block content %}
<h1>{{ database }}</h1>
<p><a href="/{{ database }}-{{ database_hash }}.db">download {{ database }}.db</a></p>
<style>
td {
vertical-align: top;
border-top: 1px solid #666;
padding: 2px 4px;
}
</style>
<table>
<tr>
{% for column in columns %}<th scope="col">{{ column }}</th>{% endfor %}
</tr>
{% for row in rows %}
<tr>
{% for td in row %}
<td>
{% if loop.index == 2 and row.type in ("table", "view") %}
<a href="/{{ database }}-{{ database_hash }}/{{ td }}">{{ td }}</a>
{% else %}
{{ td }}
{% endif %}
</td>
{% endfor %}
</tr>
{% endfor %}
</table>
{% endblock %}

View file

@ -0,0 +1,13 @@
{% extends "base.html" %}
{% block title %}Databases{% endblock %}
{% block content %}
<h1>Database{% if databases|length != 1 %}s{% endif %}</h1>
{% for database in databases %}
<h2><a href="{{ database.path }}">{{ database.name }}</a></h2>
<p>{{ "{:,}".format(database.total_rows) }} rows in {{ database.tables_count }} table{% if database.tables_count != 1 %}s{% endif %}</p>
<p>{% for table, count in database.tables_truncated %}<a href="{{ database.path }}/{{ table }}" title="{{ count }} rows">{{ table }}</a>{% if not loop.last %}, {% endif %}{% endfor %}{% if database.tables_more %}, <a href="{{ database.path }}">...</a>{% endif %}</p>
{% endfor %}
{% endblock %}

View file

@ -0,0 +1,35 @@
{% extends "base.html" %}
{% block title %}{{ database }}: {{ table }}{% endblock %}
{% block content %}
<h1><a href="/{{ database }}-{{ database_hash }}">{{ database }}</a></h1>
<h2><a href="/{{ database }}-{{ database_hash }}/{{ table }}">{{ table }}</a></h2>
<style>
td {
white-space: pre;
vertical-align: top;
border-top: 1px solid #666;
padding: 2px 4px;
}
</style>
<table>
<tr>
{% if primary_keys and row_link %}<th scope="col">Link</th>{% endif %}
{% for column in columns %}<th scope="col">{{ column }}</th>{% endfor %}
</tr>
{% for row in rows %}
<tr>
{% if primary_keys and row_link %}
<td><a href="/{{ database }}-{{ database_hash }}/{{ table }}/{{ row_link(row) }}">{{ row_link(row) }}</a></td>
{% endif %}
{% for td in row %}
<td>{{ td }}</td>
{% endfor %}
</tr>
{% endfor %}
</table>
{% if took_ms %}<small>Took {{ took_ms }}</small>{% endif %}
{% endblock %}

View file

@ -0,0 +1,35 @@
{% extends "base.html" %}
{% block title %}{{ database }}: {{ table }}{% endblock %}
{% block content %}
<h1><a href="/{{ database }}-{{ database_hash }}">{{ database }}</a></h1>
<h2>{{ table }}{% if total_rows != None %} ({{ "{:,}".format(total_rows) }} total row{% if total_rows == 1 %}{% else %}s{% endif %} in this table){% endif %}</h2>
<style>
td {
white-space: pre;
vertical-align: top;
border-top: 1px solid #666;
padding: 2px 4px;
}
</style>
<table>
<tr>
{% if primary_keys and row_link %}<th scope="col">Link</th>{% endif %}
{% for column in columns %}<th scope="col">{{ column }}</th>{% endfor %}
</tr>
{% for row in rows %}
<tr>
{% if primary_keys and row_link %}
<td><a href="/{{ database }}-{{ database_hash }}/{{ table }}/{{ row_link(row) }}">{{ row_link(row) }}</a></td>
{% endif %}
{% for td in row %}
<td>{{ td }}</td>
{% endfor %}
</tr>
{% endfor %}
</table>
{% if took_ms %}<small>Took {{ took_ms }}</small>{% endif %}
{% endblock %}