Windows>> Cymantix>> 返回
项目作者： vincentlux

项目描述 ：
Pilot study on email retrieval system
高级语言： Python
项目地址: git://github.com/vincentlux/Cymantix.git
创建时间： 2018-10-15T18:10:18Z
项目社区：https://github.com/vincentlux/Cymantix
开源协议：
下载


Cymantix

Pilot study on email retrieval system

Run

  • git clone https://github.com/vincentlux/Cymantix.git
  • pip install -r requirements.txt
  • python cymantix_grammar.py —command ‘command here’

Part1: Grammar

EBNF (with regexes):

  1.     all             ::= space op space sc
  2.     sc              ::= sc_EMAIL_from | sc_EMAIL_attach | sc_EMAIL_piece | space
  3.     sc_EMAIL_piece  ::= "EMAIL last" space op_LAST_piece
  4.     sc_EMAIL_from   ::= "EMAIL from" space op_lit_name
  5.     sc_EMAIL_attach ::= "EMAIL" space sc_attach
  6.     sc_attach       ::= "MSWORD" | "PDF" | "GIF"
  7.     op              ::= op_trig space op_first space
  8.     op_first        ::= op_TOTAL | op_LAST | op_lit_ON | op_lit_name 
  9.     op_lit_ON       ::= op_lit_name* op_ON space op_lit_topic space sc_attach* space op_LAST*
  10.     op_lit_topic    ::= "'" chars "'"
  11.     op_lit_name     ::= ("'" (chars space)+ "'" space)+
  12.     op_ON           ::= "ON"
  13.     op_LAST         ::= "LAST" space (op_LAST_time | op_LAST_piece)
  14.     op_LAST_piece   ::= ~"[0-9]*" 
  15.     op_LAST_time    ::= ~"[0-9]*" space ~"[a-z]+"
  16.     op_TOTAL        ::= "TOTAL"
  17.     op_trig         ::= "?"
  18.     space           ::= " "*
  19.     chars           ::= ~"[A-z0-9]*"

Done:

  • Literature review and base language selection
  • Finish ?TOTAL, “name” (1025)
  • Test example of NodeVisitor (1026)
  • Finish SimpleEvaluator(parsimonious.NodeVisitor) to test (1026, 1029)
  • Finish ?TOTAL, ON “topic”, “LAST”, “name” (1029)
  • Finish main Cymantix grammar

Todo:

  • Debug:
    • Add double quote compatibility
    • Date issue
    • Instance vs Class issue
  • Convert email to xml
  • Web-based retrieval system

Design notes

  • Desired result

    • Save all emails in MongoDB

    • Input:

      • Json file, each one represents an email (key-value)

      • Cymantix Command

        ?”Mike” ON “picnic” MSWORD LAST

    • Output:

      • Parsed Cymantix command

        {'from': "'Mike' ", 'topic': "'Soccer'", 'attachment': 'MSWORD', 'piece': '1'}

      • Use parsed Cymantix command to retrieve corresponding emails

Part2: Web app

  • Technologies:
    • Vue.js; Python Flask; Solr

Todo:

  • Add voice (0131-0201)
  • Move miksa3 Solr indexed item to own server(0131)
  • Optimize webpage for refs and hints
  • simple.unc.edu

Done:

  • Node.js workflow (1105-1111)
  • Solr installation on vm (1112)
  • Finish converting all emails to xml (1114-1118)
  • Escape xml field in mail2xml.py to avoid solr xml bug (0111)
  • change-schema.sh(0111)
  • Add personalized schema(0115)
  • escaped all files on the server(0115)
  • time conversion (in mail2xml.py) (0116)
  • successfully query latest date
  • Modify Cymantix output grammar
  • Research on Solr search grammar
  • Extract attachment information( or just change search query) (0116-0117)
  • cymantix_solr function to convert cymantix grammar to solr grammar
  • JCDL submitted

Schema API

Difference between StrField and TextField

Working with dates