MTE Database Schema

1. targets: listing of unique target names

   Column	Content
   target_id	unique target id: target_name + '-' + mission
   target_name	target name (each word capitalized, underscores between words)
   mission	mission to which this target belongs ("mpf", "phx", "spirit", "opportunity", or "msl")

2. aliases: mapping of target names as they appear in documents to a canonical name, to allow unification of results for all naming variants of a given target.

   Column	Content
   target_name	target name (each word capitalized, underscores between words)
   canonical_name	canonical target name (spelling corrected)

3. components: canonical listing of unique component names

   Column	Content
   component_name	canonical component name (each word capitalized)
   component_label	"Element" or "Mineral"

4. properties: canonical listing of unique property names

   Column	Content
   property_name	canonical property name (all lower case, spaces between words)

5. documents: each document in the collection

   Column	Content
   doc_id	id of source document: consists of year + '_' + abstract id
   abstract	abstract id (integer)
   title	title
   authors	authors (currently all one string)
   primary_author	primary author name
   affiliations	author affiliations (currently all one string)
   year	year of publication
   venue	publication venue, for citation (e.g., "Lunar and Planetary Science Conference, Abstract #2591")
   doc_url	source URL (e.g., PDF)
   content	raw text
   reviewer	name of reviewer, or empty if not reviewed

6. contains: table linking targets to components

   Column	Content
   target_id	unique target id: target_name + '-' + mission
   component_name	canonical component name
   target_sentence_id	sentence identifier: consists of doc_id + '-' + sentence_number; locates the target
   component_sentence_id	sentence identifier: consists of doc_id + '-' + sentence_number; locates the component

7. has_property: table linking targets to properties

   Column	Content
   target_id	unique target id: target_name + '-' + mission
   property_name	canonical property name
   target_sentence_id	sentence identifier: consists of doc_id + '-' + sentence_number; locates the target
   property_sentence_id	sentence identifier: consists of doc_id + '-' + sentence_number; locates the property

8. mentions: table linking targets to sentences in which they appear

   Column	Content
   target_id	unique target id: target_name + '-' + mission
   sentence_id	sentence identifier: consists of doc_id + '-' + sentence_number

9. sentences: table of sentences used by contains and mentions

   Column	Content
   doc_id	id of source document: consists of year + '_' + abstract id
   sentence_id	sentence identifier: consists of doc_id + '-' + sentence_number
   sentence	text of the sentence