Hydrogens disappear in roundtrip reactions

I'm trying to do a simple test of my Reactor (and my writing of reactions), by round tripping a reaction.  

I wrote a simple SMARTS reaction to reduce carbonyls with LiAlH4:

([C:1](=O)[OH]).([AlH4]

• )>>[CH2:1][OH]
  
  

  and I'm trying to convert CC(=O)O and CCO back and forth.  My definition of Lithium Aluminum Hydride is [AlH4]

• .
  
  

   

  
  

  First, I run a retro synthetic reaction from the ethyl alcohol and get back: [H]OC(=O)C([H])([H])[H] and

• [Al].  Well, the acid looks fine but the Lithium Aluminum Hydride has lost its hydrogen.  When I try to feed these back into the reaction in the forward direction the
• [Al] does not match the [AlH4]
• in the reaction definition.   I've tried to fix the hydrogen loss by adding explicit hydrogen, as you can see.  Is it better to remove explicit hydrogen from all reactions instead?  How do I know that will preserve the semantics?
  
  

   

Our SMARTS definiton of LAH is:

• .[#13;v4-]
  
  

  However, I recommend writing the reaction definition as follows:

  
  

  ([O:3][C:1]=[O:2])>

• .[#13;v4-]>[C:1][O:2]
  
  

  
  In this case, LAH 'appears on the reaction arrow', it won't be involved in the reaction in Reactor's point of view.

  
  

  The reaction can be run as:
  react -r "([O:3][C:1]=[O:2])>

• .[#13;v4-]>[C:1][O:2]" "CC(=O)O"
  
  or the reverse reaction:
  
  react -r "([O:3][C:1]=[O:2])>
• .[#13;v4-]>[C:1][O:2]" -s "CCO"
  
  

   

  
  

  Best regards,

  
  

  Viktoria

Generally, when we write reactions we should not specify hydrogens, as you've done in your example?

You do not need to specify hydrogens.