length-constrained maximum-sum subtree algorithms

• "= \ " to ""
• "\ " to " "
• "=\r\n" to ""
• "\r\n=" to ""
• "[IMAGE]" to ""
• "\r\n" to " "
• "\n" to " " # collapse to one line

Def of word:

1. only alphabetic letters: 13041 unique tokens
2. otherwise: 21862 unique tokens

1. davis utilities san plant plants times million utility blackouts generators commission customers trading companies percent electric officials federal wed edison California eletricity crisis
2. ect iso enronxgate amto confidential report draft enroncc susan joe communications ken comments order david june transmission markets language chairman Ken's email
3. bush jones president dow stock bank companies trading dynegy confidential news service natural oil credit services copyright deal percent policies Bush and Ken Lay: Slip Slidin' Away
4. davis utilities edison billion federal generators utility commission governor plan million crisis san plants electric pay companies thursday iso southern. Davis buy transmission lines

• Enron Crisis Timeline

should contain fields:

• message_id
• subject
• body
• timestamp
• datetime(optional)
• sender_id
• recipient_ids

For the last two fields, it can be replaced by participant_ids

gen_cand_trees.sh and check_events.sh

• B_on_algos.sh: increaing budget vs treesize on various algorithms