Identifying relevant content in email
View Patent ↗Message management services can include processing an email to identify relevant content from among all content in the email and converting the identified relevant content into a simple, easy-to-read format. For example, message management services can apply multiple parsing strategies to an email. Each strategy can attempt to parse the email to identify relevant content and to output results that include any identified relevant content and an associated confidence score. The results having the highest confidence score can be selected. The identified relevant content included in the selected results can be converted into a flat string with stylings and links for displaying relevant content of the email in a simple, easy-to-read format.
1. A method comprising:
processing, using a plurality of strategy modules implemented on a server system, an email to identify relevant content included in the email, where the plurality of strategy modules include a first strategy module for identifying relevant content in emails from a first message provider and a second strategy module for identifying relevant content in emails from a second message provider;
generating, by the first strategy module, a first identified relevant content and a first corresponding confidence score, where the first confidence score is based on a first pattern that corresponds to a first type of message boundary that separates a reply email from a previous email in a thread;
generating, by the second strategy module, a second identified relevant content and a second corresponding confidence score, where the second confidence score is based on a second pattern that corresponds to a second type of message boundary that separates the reply email from the previous email in the thread;
comparing, by a strategy manager, the first corresponding confidence score and the second corresponding confidence score;
selecting, by the strategy manager, the first identified relevant content as a selected identified relevant content when the first corresponding confidence score is greater than the second corresponding confidence score based on the comparison, where each confidence score is based on how closely the pattern detected in the email matches the message boundary pattern that the plurality of strategy modules can be configured to detect; and
displaying the selected identified relevant content that has the highest corresponding confidence score.
2. The method of claim 1 , wherein the email is a latest email in a thread and includes contents of a previous email.
3. The method of claim 2 , wherein, for each of the strategy modules, processing the email further includes:
detecting in the latest email a pattern that closely matches a message boundary pattern, the message boundary pattern indicating a message boundary in the latest email, the message boundary separating contents of the latest email and the contents the previous email in the thread; and
identifying a latest contribution in the contents of the latest email, the latest contribution being relevant content in the contents of the latest email.
4. The method of claim 3 , wherein, for each of the strategy modules, processing the email further includes:
constructing a tree structure having a plurality of nodes that correspond to the latest email and the previous email; and
placing the relevant contents of the previous email and the latest email into the corresponding nodes of the tree structure.
5. The method of claim 1 , further comprising:
obtaining, by a native converter implemented on a computer, the relevant content selected by the strategy manager; and
converting, by the native converter, the relevant content selected by the strategy manager to a native format.
6. The method of claim 5 , wherein the native converter converts the relevant content selected by the strategy manager to the native format by:
identifying a base node from among the nodes of the tree structure; converting the relevant content of the base node to the native format; recursively walking to a next node of the nodes of the tree structure; and converting the relevant content of the next node to the native format.
7. The method of claim 6 , wherein the native converter is configured to continue walking the tree structure until the relevant content has been converted to the native format.
8. The method of claim 5 , wherein the native format is a flat string with a plurality of stylings.
9. The method of claim 5 , wherein the native converter converts the relevant content selected by the strategy manager from a HTML DOM (Document Object Model) to a flat string with a plurality of stylings.
10. A server system comprising:
a strategy manager server having a processor and a memory;
a plurality of strategy modules stored in the memory of the strategy manager server, where the plurality of strategy modules include a first strategy module for identifying relevant content in emails from a first message provider based on an email format used by the first message provider and a second strategy module for identifying relevant content in emails from a second message provider based on an email format used by the second message provider, each of the strategy modules being configured to:
detect in a latest email a pattern that closely matches a message-boundary pattern, the message-boundary pattern indicating a message boundary in the latest email, the message boundary separating contents of the latest email and the contents the previous email in the thread;
identify a latest contribution in the contents of the latest email, the latest contribution being relevant content in the contents of the latest email; and
determine a confidence score indicative of how closely the pattern detected in the latest email matches the message-boundary pattern; the strategy manager server being configured to:
apply two or more of the strategy modules to the latest email; compare two or more confidence scores; and select from among two or more latest contributions identified by the strategy modules a latest contribution associated with a highest confidence score based on the comparison.
11. The server system of claim 10 , wherein the latest contribution is part of a conversation between a sender of the previous email and the sender of the latest email.
12. The server system of claim 11 , wherein each of the strategy modules include program code to:
construct a tree structure having first and second nodes that correspond to the previous email and the latest email;
place relevant content of the previous email in the first node of the tree structure; and place the relevant content of the latest email in the second node of the tree structure.
13. The server system of claim 12 , further comprising:
a native converter server having a processor and a memory, the native converter being configured to:
identify a base node from among the first and second nodes of the tree structure;
convert the relevant content of the base node to the native format; recursively walk to a next node of the tree structure; and convert the relevant content of the next node to the native format.
14. The method of claim 13 , wherein the native format is a flat string with a plurality of stylings.
15. A method comprising:
applying, by a server system, to an email a plurality of strategy modules, where the plurality of strategy modules include a first strategy module for identifying relevant content in emails from a first message provider based on an email format used by the first message provider and a second strategy module for identifying relevant content in emails from a second message provider based on an email format used by the second message provider;
each of the strategy modules being configured to:
detect in the email a pattern that closely matches a message-boundary pattern, the message-boundary pattern being indicative of a message boundary that separates a latest email from a previous email in the thread;
identify relevant content of latest email as the latest contribution; and determine a confidence score indicative of how closely the pattern detected in the email matches the message-boundary pattern; and
comparing two or more confidence scores;
selecting, by the server system, from among one or more of latest contributions identified by two or more of the strategy modules a latest contribution associated with a highest confidence score based on the comparison.
16. The method of claim 15 , wherein the strategy modules are further configured to construct a conversation by:
constructing a tree structure having a plurality of nodes that correspond to the previous email and latest email; and
placing relevant contents of the previous email and the selected latest contribution of the latest email in the corresponding nodes of the tree structure.
17. The method of claim 16 , further including:
converting, by the server system, the relevant contents of the previous email and the selected latest contribution of the latest email.
18. The method of claim 17 , wherein the native format is a flat string with a plurality of stylings.
19. A non-transitory computer-readable storage medium having stored thereon executable instructions that, when executed by one or more processors of a computer system, cause the computer system to execute a method comprising:
applying to an email a plurality of strategy modules, where the plurality of strategy modules include a first strategy module for identifying relevant content in emails from a first message provider based on an email format used by the first message provider and a second strategy module for identifying relevant content in emails from a second message provider based on an email format used by the second message provider, each of the strategy modules being configured to:
detect in the email a pattern that closely matches a message-boundary pattern, the message-boundary pattern being indicative of a message boundary that separates a latest email from a previous email;
identify relevant content of the latest email as a latest contribution in the email; and
determine a confidence score for the latest contribution, the confidence score being indicative of how closely the pattern detected in the email matches the message-boundary pattern; and
compare two or more confidence scores;
selecting from among one or more of latest contributions identified by two or more of the strategy modules a latest contribution having a highest confidence score based on the comparison.
20. The computer-readable storage medium of claim 19 , wherein the strategy modules are further configured to construct a conversation by:
constructing a tree structure having a plurality of nodes that correspond to the previous email and the latest email; and
placing relevant content of the previous email and latest email into the corresponding nodes of the tree structure.
21. The computer-readable storage medium of claim 19 , wherein the method further includes:
obtaining the latest contribution selected by the strategy manager; and converting the latest contribution to a native format.
22. The computer-readable storage medium of claim 21 , wherein the native format is a flat string with a plurality of stylings.
23. A server system, comprising:
a strategy manager server having a processor and a memory;
a plurality of strategy modules stored in the memory of the strategy manager server, where the plurality of strategy modules include a first strategy module for identifying relevant content in emails from a first message provider based on an email format used by the first message provider and a second strategy module for identifying relevant content in emails from a second message provider based on an email format used by the second message provider, the strategy modules configured to:
detect in an email a format that closely matches at least one of a latest reply or a latest forward;
detect in the email a pattern that closely matches a message-boundary pattern, the message-boundary pattern being indicative of a message boundary that separates at least one of a latest reply or a latest forward from a previous email;
identify relevant content of one of the latest reply or the latest forward as a latest contribution of the email; and
determine a confidence score for the latest contribution, the confidence score being indicative of how closely the pattern detected in the email matches the message-boundary pattern;
the strategy manager server being configured to:
apply some or all of the strategy modules to an email; compare two or more confidence scores; and select from among one or more latest contributions identified by some or all of the strategy modules a latest contribution having a highest confidence score based on the comparison.
24. The server system of claim 23 , wherein the strategy modules are further configured to construct a conversation by:
constructing a tree structure having a plurality of nodes that corresponds to a previous email and at least one of a latest reply or a latest forward; and
placing relevant content of the previous email and the latest contribution of one of the latest reply or latest forward into the corresponding nodes of the tree structure.
25. The server system of claim 23 , further including: a native converter server having a processor and a memory, the native converter being configured to convert the latest contribution selected by the strategy manager to a native format.