<div dir="ltr">Hi Megha,<div><br></div><div>I'm the main developer of PyDelphin. It is not the canonical tool for MRS-DMRS conversion (that is the LKB), but I think it might be the most commonly used for the task.</div><div><br></div><div>Install it like this (commands begin after $):</div><div><br></div><div> $ pip install pydelphin</div><div><br></div><div>Then you can convert (Simple)MRS to DMRS like this:</div><div><br></div><div> $ delphin convert --from simplemrs --to dmrx every-student-reads-a-book.mrs </div><div> <dmrs-list><dmrs cfrom="-1" cto="-1">...</dmrs></dmrs-list></div><div><br></div><div>You can convert from DMRS to MRS like this:</div><div><br></div><div> $ delphin convert --from dmrx --to simplemrs every-student-reads-a-book.xml</div><div> [ TOP: h0 RELS: <...> HCONS: <...>]</div><div><br></div><div>Details on conversion:</div><div><br></div><div>In well-formed and algebra-compliant MRSs (i.e. from a well-behaved grammar), conversion to and from DMRS should be lossless. Note that losslessness does not imply string equivalency, but the meaning will be the same.</div><div><br></div><div>In MRS<->DMRS conversion, the following is a partial list of situations I've seen that can cause problems:</div><div><br></div><div>* 2+ elementary predications (EPs) sharing an intrinsic variable (ARG0)</div><div>* EPs without an ARG0</div><div>* scoping problems (e.g. an EP is qeq'd to its own label)</div><div>* disconnected MRSs (some EP does not have an argument, or is not the argument of, or does not share a scope (i.e. label) with any other EP in the graph starting at TOP)</div><div>* Coindexation of unexpressed EPs in MRS (e.g. i-variables standing for dropped arguments) cannot be recovered when round-tripping to DMRS<br></div><div><br></div><div>These are simply bad MRSs, and we don't guarantee to produce an equally bad DMRS.</div><div><br></div><div>Also, individual constraints (ICONS) are a relatively new feature in MRSs, and we have not yet implemented a solution in DMRS, so those will be lost.</div><div><br></div><div>There are some other edge cases that never come up in practice.</div><div><br></div><div>I hope this was useful and informative. Please let me know if anything is unclear.</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Feb 20, 2017 at 9:05 AM, Woodley Packard <span dir="ltr"><<a href="mailto:sweaglesw@sweaglesw.org" target="_blank">sweaglesw@sweaglesw.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><br> Megha,<br> <br> I believe PyDelphin is the tool you will want to use for that conversion. I am forwarding your inquiry to the Delphin developers list, since my own tools and expertise do not directly bear on your question. I believe there are people on list who can help though. Can you give examples of the DMRSs you are having trouble with and let us know where they came from?<br> <br> Thanks, Woodley<br> <br> <br> > On Feb 20, 2017, at 2:50 AM, megha jain <<a href="mailto:jain11megha@gmail.com">jain11megha@gmail.com</a>> wrote:<br> ><br> > Hello.<br> ><br> > I want to ask details about conversion from MRS to DMRS and DMRS to MRS.<br> ><br> > In many sentences it is easy to convert from MRS to DMRS from the package already available but problem comes while DMRS to MRS conversion.<br> ><br> > Can you let me know the approach comes in use or code which handles it.<br> ><br> > I dropped a mail to Berthold Cryssman as you suggested me but couldn't get any response till now.<br> ><br> > Please do respond.<br> ><br> > Thank You.<br> <br> </blockquote></div><br><br clear="all"><div><br></div>-- <br><div class="gmail_signature" data-smartmail="gmail_signature"><div dir="ltr">Michael Wayne Goodman<div>Ph.D. Candidate, UW Linguistics</div></div></div> </div>