def new_commit(self, content, pdfbits):
"""
Handle new commit based on [content] and [pdfbits].
inputs:
content: Contents of the paper.
pdfbits: The pdf file in bits.
"""
# Check all required field exist
title = content['title']
keywords = content['keywords']
timestamp = content['timestamp']
descriptions = content['descriptions']
# Write pdf file to [fname]
# Get legal file name
fname = self._title2fname_(title)
with open(os.path.join(self.papers_dir, fname), 'wb') as f:
f.write(pdfbits)
# Update keywords
for kw in keywords:
print(kw)
self.insert_keyword(kw, title, timestamp)
# Update descriptions
for desc in descriptions:
print(desc)
self.insert_description(desc, descriptions[desc], title)
# Done
logger.info(f'PAPERS_SERVER received new commit of {title}')
def update(self):
"""
Update pdfs from buffer folder.
yield:
self.pdfs
"""
self.read_ignores()
pdfs = pd.DataFrame()
# Walk through folder
# names = [n for n in os.listdir(self.buffer_dir)
# if all([n.endswith('.pdf'),
# n not in self.ignores.name.values])]
names = [n for n in os.listdir(self.buffer_dir) if n.endswith('.pdf')]
paths = [os.path.join(self.buffer_dir, n) for n in names]
pdfs['name'] = names
pdfs['path'] = paths
# Parse informations
for entry, method in [('atime', os.path.getatime),
('ctime', os.path.getctime),
('mtime', os.path.getmtime)]:
pdfs[entry] = [
time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(method(p)))
for p in paths
]
# Set index
self.pdfs = pdfs.set_index('name', drop=True)
# Done
logger.info('Buffered file names updated.')
def __init__(self):
self.papers_dir = profiles.papers_dir
self.keywords_path = os.path.join(self.papers_dir, 'keywords.json')
self.descriptions_path = os.path.join(
self.papers_dir, 'descriptions.xlsx')
self.read_keywords()
self.read_descriptions()
logger.info('PAPERS_SERVER started.')
def new_ignore(self, name):
"""
Add new ignores by [name].
Update this buffer.
inputs:
name: File name to be ignored.
yield:
Update self.ignores and write to disk.
"""
if name not in self.ignores.name.values:
self.ignores = self.ignores.append({'name': name},
ignore_index=True)
self.ignores.to_json(self.ignores_path)
logger.info(f'Ignore name: {name}')
self.update()
def get_by_name(self, name):
""" Get paper and its path based on [name]
output:
bits: Bitstream of the pdf file
fpath: Path to the pdf file """
# Assert we have the paper
assert (name in self.pdfs.index)
# Get fpath
fpath = self.pdfs['path'][name]
info = PdfReader(fpath).Info
# Get bits
with open(fpath, 'rb') as f:
bits_list = f.readlines()
bits = b''.join(bits_list)
logger.info(f'BUFFER_SERVER get_by_name success on {name}')
return fpath, info, bits
def buffer_get(self, name, method='open'):
"""
Get file by [name] in buffer_server using [method='open' or 'start']
method: 'open' means return bits stream
'start' means start the file using default app
"""
try:
assert (method in ['start', 'open'])
fpath, info, bits = self.buffer_server.get_by_name(name)
if method == 'open':
return bits
if method == 'start':
logger.info(f'WORKER buffer_get starts {name}')
os.system(fpath)
return None
except Exception as e:
logger.error(f'WORKER buffer_get failed: {e}')
return None
def buffer_commit(self, name, content):
""" Handle new commit based on [name] and [content].
Return 0 if success,
return others if failed. """
# Parse [content]
try:
new_content = dict(
timestamp=float(content['date']) / 1000, # Commit timestamp
title=content['title'], # Title of the paper
keywords=[
e.strip().title() for e in content['keywords'].split(',')
if e.strip()
], # Keywords of the paper, list
descriptions=self._description2dict_(content['descriptions'])
) # Descriptions of the paper, dict
logger.info(f'WORKER buffer_commit parsed content')
print(new_content)
except Exception as e:
logger.error(
f'WORKER buffer_commit failed on parsing content: {content}, error: {e}'
)
return 1
# Get pdfbits
pdfbits = self.buffer_get(name)
if pdfbits is None:
logger.error(
f'WORKER buffer_commit failed on getting pdf file {name}')
return 1
try:
# Commit to papers_server
self.papers_server.new_commit(new_content, pdfbits)
# Ignore new name in buffer_server
self.buffer_server.new_ignore(name)
logger.info(f'WORKER buffer_commit committed {new_content}.')
return 0
except Exception as e:
logger.error(
f'Worker buffer_commit failed on committing content: {new_content}, error: {e}'
)
return 1
def get_by_title(self, title, fields):
""" Get paper and its content by [title] according to [fields].
inputs:
title: Title of the paper
fields: Fields to be returned
outputs:
Outputs are according to [fields], may contains following fields
fpath: Path to the pdf file
bits: Bits stream of pdf file
keywords: Keywords
descriptions: Descriptions """
contents = dict()
# Get fname and fpath
fname = self._title2fname_(title)
fpath = os.path.join(self.papers_dir, fname)
if 'fpath' in fields:
contents['fpath'] = fpath
# Assert descriptions, keywords and pdf file exists
assert(title in self.descriptions.index)
assert(title in self.keywords.index)
assert(os.path.exists(fpath))
# Get contents
# Descriptions
if 'descriptions' in fields:
descriptions = self.descriptions.loc[title]
contents['descriptions'] = descriptions[descriptions.notna()]
# Keywords
if 'keywords' in fields:
keywords = self.keywords.loc[title]
contents['keywords'] = keywords[keywords.notna()]
# Get bits of pdf file
if 'bits' in fields:
with open(fpath, 'rb') as f:
bits_list = f.readlines()
contents['bits'] = b''.join(bits_list)
logger.info(f'PAPERS_SERVER get_by_title success on {title}')
return contents
def run(self, ip='localhost', port=8612):
""" Run socket listening on [ip]:[port] """
# Setup socket listener
sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
sock.bind((ip, port))
sock.listen(1)
logger.info(f'WEBSERVER listen on {ip}:{port}')
# Serving
idx = 0
while self.running:
# Accept new connection
connection, client_address = sock.accept()
logger.info(
f'WEBSERVER is connected {connection} from {client_address}.')
# Start new thread to serve the connection
t = threading.Thread(target=self.serve_connection,
args=(connection, client_address, idx))
t.start()
# idx increase
idx = (idx + 1) % 65536
logger.info("WEBSERVER stopped.")
def serve_connection(self, connection, address, idx=None):
"""
Method to serve [connection] of [idx] from [address].
"""
try:
# Fetch request
request = connection.recv(65536).decode()
length = len(request)
logger.info(f'WEBSERVER-{idx} receives {length} bits')
# Respond
content = self.respond(request)
if not isinstance(content, bytes):
content = content.encode()
length = len(content)
logger.info(f'WEBSERVER-{idx} responses {length} bits')
connection.sendall(content)
idx += 1
except Exception as e:
logger.error(
f'WEBSERVER runtime error. connection={connection}, client_address={address}, error={e}'
)
finally:
connection.close()
logger.info(f'WEBSERVER-{idx} connection closed')
def __init__(self,
buffer_server=BUFFER_SERVER(),
papers_server=PAPERS_SERVER()):
self.buffer_server = buffer_server
self.papers_server = papers_server
logger.info('WORKER initialized.')
def __init__(self):
self.buffer_dir = profiles.buffer_dir
self.ignores_path = os.path.join(self.buffer_dir, 'ignores.json')
self.update()
logger.info('BUFFER_SERVER started.')
def __init__(self, ip='localhost', port=8612):
self.running = True
logger.info("WEBSERVER initialized.")
